Does Mechanisation and automation actually reduce the number of labourers overall?

In My Humble Opinion
21

There is currently a large shortage of factory workers. The unemployment rate is at record lows, despite the weak economy.

There is zero evidence that automation reduces overall jobs, and reams of evidence that it doesn’t.

Automation is generally not done to save labor. Mechanization was labor-saving, which is one reason why agricultural output exploded in the 20th century while food prices went down and the number of people involved in the food industry went down.

Today, factories are automating for three reasons:

  1. To keep up with supply. If you are going to make 100 million iPhones per year, you’re going to have to automate production.

  2. Quality. Humans add a lot of variance to manufacturing, and variance is the enemy of quality. Modern factories thrive on known, documented processes and exact repeatability. When you are looking to shave 1/1000 off of a dimension to save material and energy, you need machines that repeat a process exactly, every time.

  3. Lack of workers. Factories are digitizing their processes and automating because their workforces are shrinking and they don’t have apprentices to learn how to do things from senior workers. So once those workers retire they are taking domain knowledge with them. For whatever reason, young people don’t want to work in factories, despite the high pay and in many, very good working conditions and benefits.

Automation is very expensive. It’s not done just do you can lose that guy on the line who threads a bolt. We used to sell maintenance contracts for $200,000 per year just for our automation software. For large corporations, it was much more. But you can’t get super high quality goods with maximum efficiency and volume without automating.

But the reason this doesn’t cost jobs is because humans are valuable, and a human freed from threading bolts onto a spindle all day is a human available for higher-value work, and the economic gains we get from automation allow markets to expand, which creates more jobs overall. This has been true since we started producing things. There’s no reason to believe that will change.

22

Another point to consider: Automation definitely reduces the number of man-hours required, but that doesn’t necessarily mean a reduction in the number of jobs. You could keep the same number of workers, with each of them working fewer hours. Which might mean lower pay, but that’s not a problem if everything the workers want to buy is cheaper, due to automation in those industries. You could end up with everyone still getting the things they want and need, and just doing less work for it.

23

Except there is reason to believe that will change; in the past (the supposed “Second Industrial Revolution” if you think the effects of innovation can be cleanly broken into epochs) mechanization and standardization largely replaced individual craft skill, supplanting it with machinery that could produce consistent parts and a more-or-less universal system of technical education that taught people how to be precise functional units rather than the prior system of apprenticeship to learn skills from an individual master for many years before advancing to independent proficiency. This both allowed for higher rates of production but also, like the Agricultural Revolution before it, created enough leisure time to free people to engage in intellectual pursuits, universal literacy, and education that was not immediately applicable to skilled trades was accessible by more than a tiny elite.

The difference today is that a lot of focus on ‘automation’ is shifting the burden of intellectual labor to machine systems that are now capable (if incompletely and sometimes unreliably) of decision-making, planning, interpretation, et cetera using fairly high level algorithmic definitions and rules and then ‘learning’ to adapt them to novel situations. (Whether this adaptation is true ‘learning’ or not is a worthwhile but mostly semantic argument; it is clear that machine intelligence is as adaptable as human intellect even though it really does not function the same way, and the problem of making it reliable is more in being able to regulate and guide the rationalization process to decisions and outcomes that are consistent with our own values and needs.) This shift in intellectual labor—something formerly the exclusive domain of human workers—is now literally taking away jobs in the expanding technology sectors that would otherwise call for more human labor without producing additional work that only humans can do. For the first time in human history, human aren’t just being displaced as physical labor or to perform ‘skilled’ physical work but are actually losing on an intellectual basis to automation.

This is not to say that human labor is about to become obsolete; despite Elon Musk’s dancing ‘android’, we are nowhere near robots that can reliably perform any but a small subset of household tasks, and even if we could make robots capable of physically caring for children, the elderly, and disabled, there is a strong argument that we should not become reliant upon this for anything but assistance because of the value of social contact that the ‘human touch’ brings, particularly to people who are often in greatest need of it. Nor will we have ‘teaching machines’ to replace the broad array of things that teachers do beyond simply presenting information and testing students, or robot standup comedians, or therapists, or any of a vast number of professions that ostensibly could be taken over by sophisticated algorithms but would lose essential qualities by doing so.

But many jobs—particularly ‘junior level’ tasks like doing literature searches, or writing grant proposals, or generating documentation—could be largely automated with coming improvements in machine intelligence with greater reliability and lower cost, essentially eliminating those roles in law practice, academia, engineering, et cetera, and the effects of this would certainly be to displace people who would otherwise be filling these roles rather than creating new fields of endeavor for which only human intellect is adequate. As autonomously piloted vehicles will inevitably replace taxi and rideshare drivers (although not, I think, and universally or nearly as soon as enthusiasts would like to imagine), machine intellect will replace many positions in a wide array of industries, limiting entry-level positions to those who demonstrate exceptional skill or creative ability (or who just have inside connections with the human decision makers) without creating more jobs in that field or attendant to it. This is a historically unique change in employment in which humans are suddenly not the smartest intellect in the room (at least, in certain capacities) and is qualitatively different than a shift from mechanization of mechanical labor or even electronic computers replacing human ‘calculators’ in doing computations faster and with more accuracy. It is a point at which computers and automated systems will start giving themselves orders rather than blindly following algorithms defined by humans and represents a distinct shift in the place that people have in the system of production, research, design, and a wide array of services.

Stranger

24

It’s not always about “overall”. Individuals matter, too. When a plant shuts down in a rural town in Indiana, the whole town suffers. The “free market” doesn’t find new work for the displaced workers, nor replace the income that made the town viable. It doesn’t pay for schools to be improved to at least ensure that the children of these displaced workers can compete in the new labor market.

Which put a lot of farmers out of work. The same land can be tended to by a fraction of those needed before. Now, what I find interesting is that these workers were partially absorbed by factories that moved out of cities into rural areas, where land and labor were cheap, and environmental laws were lax. This also brought more money to the town, increasing the number of jobs available, absorbing the rest of those displaced from the fields, and allowing the town to grow well beyond what it could entirely from agriculture. Then jobs were cut, people were laid off as automation and offshoring reduced the need for the local labor, often the plant would just shut down, cratering the local economy. This is one of the reasons we ended up with the rust belt full of angry people who feel that their way of life is being taken away from them.

Your number 1 and 3 on that list are litterally done to save labor. You don’t need to automate to keep up with supply if you can just hire more workers, and if you have a lack of workers, you are trying to reduce the needed labor to what’s available.

As far as 2, it’s context dependent. No matter how many workers you have, no matter how skilled they are, they will not be able to make an iPhone without access to a significant amount of automation that can perform tasks far more precisely than a human can. But there are many applications, even in making an iPhone, that a human can do well enough. If you are looking to shave off .1% of your material costs in order to maximize your profit, you’ve already done as much as possible to reduce your labor costs.

$200,000 a year is equivalent to 2-5 workers, depending on pay and benefits. Don’t you think that your automation reduced the need of jobs by more than that? I’m sure that they saved quite a bit of money by not having to pay workers to do that job.

Tell that to the guy who got laid off from his spindle threading job, and is now not sure how to keep his family afloat.

Most of those economic gains go to the owners of the machines, and the losses are felt by the workers displaced. IF the cost savings go into lowering prices, then consumers have more money to spend on other products, increasing the demand for them, which can increase overall jobs, assuming the the products in demand don’t come from automated industries. However, if the cost savings go into profit, then that will have far less of a jobs creation effect.

What has been true since we started producing things is that people have been left behind by progress. Automation is overall a good thing, but we shouldn’t ignore the problems that it causes. In theory, the gains we get from automation should allow us to address these problems more effectively, but instead, the problems only seem to be getting worse.

You said that a human freed from doing a menial task is valuable, so what would you be willing to do to support that valuable human? If the spindle threader comes home to his family to tell them that he got laid off, should they have fear of what comes next, or excitement over the opportunities in front of them? In most cases, it’s the former, what would you do to make it the latter?

That ex-spindle threader isn’t going to feel valuable if he takes on the only job that he’s qualified for with lower pay and benefits. Should we support him and his family as he returns to school to gain new skills, or even to just stay home and write that book he’s never had time or energy to do?

The job market is important, it’s one of the reasons why the unemployment rate is such a highly tracked and publicized figure. Governments rise and fall on that number. We always talk about how many jobs will be brought to an area by a new business, congresscritters brag about how many jobs their legislation brought to their constituents.

We are in an anomaly right now, brought on by several factors, not the least of which was a global pandemic. I don’t think it’s the new normal, and, assuming we don’t all die from a hybrid of polio and monkey pox mutated by the fallout from global nuclear war, things will return to how they have been in the past, which is that we have a number people who can’t find a viable job. There’s no reason to believe that will change.

25

The modern societal complaint against automation - certainly in America - is that we once were a country that lifted the working class into middle-class economic status by providing them well-paying factory jobs that offered essentially lifetime guarantees of work, health benefits, and pensions. These jobs no longer exist and America is failing its future.

This argument totally ignores context. In the post-WWII era, at the height of the industrial/mechanical era, the U.S. was almost alone as a high-functioning economy in the world. Employers could afford to tender unprecedented - and unsustainable - promises to workers. We call this time the 1950s, but it really lasted about a quarter-century, from 1946 to the 1973 oil crisis.

Reality crashed upon us. The rest of the world recovered, retooled with better modern factories and technologies, and took away market share everywhere. American employers saw lower-pay conditions elsewhere and rushed to take advantage. At the same time, entry-level jobs that demanded personal contact were demeaned and kept at minimal pay levels. Entry-level, service, and gig economy jobs soared but wages did not and benefits were nearly non-existent. Even so, these jobs are being automated out of existence by ATMs, self-checkouts, touch-screen order boards, online ordering, and other electronic outposts that are nothing like factory robots.

The “1950s” were a hopefully unrepeatable anomaly, even though many people insist on calling it the base norm. The 2020’s are also an anomaly, and not really because of COVID. That just hurried the future that was already coming.

I don’t see true parallels between that future and the First and Second Industrial Revolutions. Something brand new is happening. The core element of business thought no longer involves building an institution for the future, but to provide an instant service to gratify current needs. Obviously, this is not universal, but building a half-dozen giant chip factories is not the same as building a thousand giant warehouses whose goal for the future is to reduce the number of humans to as near zero as possible.

Talking just about America in this global economy is also wrong. Soon enough, a billion people around the world will be leaving their home countries after they become unlivable. They will need housing, food, education, health care, and jobs. I see no indication whatsoever that these will be available or that automation is more than a partial solution that may be more of a huge problem.

26

Richard Florida posits in his 2010 book “The Great Reset” that over the last 150 years or so there have been three periods where productivity increased beyond society’s ability to maintain demand for those products, thus leading to increasing debt/credit, speculation, bubbles, economic crashes, and eventually a retooling of living arrangements as a means to induce new demand.

The first period was in the 1870s with the mechanization of agriculture. This led to crashing wages and excess food production. Investment money was poured into railroads to move agricultural goods, leading to significant overbuilding and various financial panics. The displaced workers moved to growing industrial cities with huge demand for new factories, tenement housing, street railway systems, and all the other accoutrements of city living. This population shift from rural to urban and the enormous construction industry and new living arrangements that came from it “reset” or bumped up demand to better match the economy’s output.

By the 1920s industrialization had reached a peak, and once again productivity had increased to such an extent that there weren’t enough buyers for those products and services anymore. Thus came the Great Depression. It wasn’t until WWII that things really swung back into action, but after the war was over, there was great fear that the Depression would come right back. So the government pushed for yet another new living condition, the suburbs. Now instead of people living in apartments and townhouses, taking streetcars or subways, sharing laundry facilities and small yards, every family now needed more land, a bigger house, appliances and lawn equipment for everyone, multiple cars, and government-funded streets and highways. A less efficient living arrangement was encouraged, if not forced, on the growing middle class, which “reset” and propped up the manufacturing industries.

Come 2008 the technological revolution had also reached a sort of peak, and economic activity driven mostly by the suburban construction industry (which includes the financing thereof, consumer products to go in it, roads and vehicles to access it, etc.) came to a head due in part to overfinancialization and nearly all economic activity being based solely on that one aspect of the economy, as nearly all others were tapped out from, once again, increased productivity saturating demand. So what’s the current “reset”?

We don’t know, since it would seem we’re in the middle of it. Covid threw a monkey wrench into everything, but the real estate market was already insane before that. Younger generations don’t hold the same regard for the trappings of suburban living, especially when it comes to cars and big houses in the middle of nowhere, yet cities are having trouble absorbing new housing after decades of disinvestment and depopulation.

The key takeaway though is that when increased productivity does not confer enough wealth on the general populace to maintain demand for the goods and services they’re providing, then shit gonna happen. Automation seems poised to make this worse, unless the fruits of that automation are more equitably distributed. Unfortunately, that seems unlikely given the sort of power (im)balance of the ruling classes. Might there be some black swan event that allows for yet another “reset” in living arrangements, like a breakthrough in fusion energy, some wild new transportation tech, Skynet? We’ll see, but it seems that increased automation/mechanization/productivity doesn’t automatically improve one’s economic outlook, it’s just that external forces have so far managed to counteract them. Will there be another intervention in the future, and will it work? I don’t know.

27

yes - “productivity” is the word you’re looking for.
With higher productivity (per person) then the goods produced become cheaper. So instead of the majority working on farms just to feed themselves, they work in factories, and the 1% or so that work on farms produce enough to feed all the rest who go work in factories.

With automation, cars become cheaper - more people can afford a car, hence more cars made. Or any other product. People with disposable income will spend it on clothes, so there is a demand - and hence jobs - producing cotton goods -and now synthetics. Instead of your meager income just barely buying food and rent, it now also buys a car, TV, computer, cellular service, banking and 401K services, etc. etc. You can think of this as “I make a lot more money” or as “with a typical person’s income, I can slice it into more small pieces, spread it around a huge amount of goods and services because they are so cheap”. To provide all those services, there is work for a huge number of people.

For example - computers put typists and secretaries out of work, they are making post office letters obsolete, they put animation inbetweeners and cel washers out of a job, they’ve destroyed newspapers and magazines, the bookstore and the travel agent for starters - but stand back, look at the whole computer industry, and absolutely how many people “work in computers” one way or another.

Retraining OTOH is a whole different story. Some people naturally change careers as needed, some adapt a bit, and some can never. But - we have one of the lowest unemployment rates in the last half-century, and the poorest of us still have toys (like smart phones) that were impossible to anyone 100 years ago, no matter how rich.

28

This presupposes that humans will continue to be the ones producing cotton goods, cars, TVs, computers, maintaining cell towers, and so on. But what if they’re not? The mobility, dexterity, and visual/auditory processing capability of robots is growing by leaps and bounds. There’s already a lot of automation in agriculture and meat-packing, but some of those tasks are stubbornly resistant to automation, or at least they have been. But we’re approaching a point where robots with AI/machine vision will be able to butcher a cow, pick apples, or re-roof a house more economically than a human worker. The people working in these kinds of jobs are already receiving next to nothing for their labor. When the robots come for their jobs, these people can’t afford to go to (or send their kids to) any sort of trade school, and even if they could, there is almost no kind of job could they train for that would be assuredly safe from automation.

29

Well, you are both right and wrong. The human is freed from threading bolts but, before automation, that was the highest valued work available for that person. Whatever job she gets will be valued at something less than threading bolts was valued at before the bolt-threading robot. It will be valued at more than threading bolts only because the bolt-threading robot made threading bolts so cheap.

I agree that the economic gains from automation do expand jobs overall.

I’m not so sure. There are jobs that robots are starting to do that many people thought robots might never be able to manage (like customer service chatbots).

I feel confident that humans will find new work, no matter the state of the economy. As the value of human labor drops, there will always be some roles where humans have a relative advantage to capital, and humans will slot themselves into those roles. There is no limit on how many goods or services humans will want to consume, so as long as we have people with some time and ability to produce goods and services, the economy will find some way to employ them providing those goods and services.

30

I disagree with the 2008 situation being similar to the earlier ones.

Since the 1980s there’s been a large surplus of investment funds with a lack of corresponding sound financial places to park it (esp. things that fundamentally grow the economy).

So in the late-ish 80s we had a real estate bubble with accompanying S&L fraud to help perpetuate it. In the late 90s we had the Internet/tech and stocks overall bubble. This soon lead into another real estate bubble (mostly residential this time). We are now in another bubble that combines the best of both: tech stocks are overvalued and real estate values are crazy. So, no worries there, right?

Of course the idiotic “solution” each time one of these bubbles crash is to “encourage investment” (primarily giving rich people more money). Which just makes things worse.

As to productivity and how this affects workers, take a look at the chart in this papers. Since 1980 productivity has increased (but not quite at the same rate as before) while wages, which used to keep up, haven’t. The money from this increased product ivy is going somewhere else.

A lot of trends like this magically seem to all start around 1980.

31

Why else are they even in business except to make profits? You say it like they’re a bad thing. And why should they be interested in providing better jobs beyond what helps those workers be more productive and/or be retained longer?

That’s the thing people seem to forget. Automation is good for businesses because the total costs over time tend to be lower, or they’re the same and accuracy/reproducibility is better (i.e. the holes are always drilled in exactly the same place, or the right amount of torque is always applied, etc…).

The big question as I see it is really whether or not the shifts in industry and the workforce end up creating other jobs elsewhere in the economy than there used to be. I mean, we’ve got more people employed than ever before, so clearly some kind of jobs were created to replace those typist, mechanical assembler, etc… jobs lost to automation in the past.

The other thing that we’ve got on the horizon is the degree/speed with which deep learning/AI will replace a lot of previously un-automatable jobs. We’re already seeing it with natural language processing coupled with deep learning for incoming call center type jobs. Something crazy like four out of five of calls to city information lines can be handled by bots these days, with only one in five needing an actual person to figure out what that person is asking, or who to send it to. That’s only going to get better as time goes on- eventually nearly all calls will be handled accurately by bots. Many other informational-type clerical jobs with a significant routine component will go that direction as well, leaving the jobs with creative, interpretation, or decision-making components for humans.

32

That may be true, but the operative phrase is “the value of human labor drops.” That means people have to work more/longer/harder to maintain their standard of living. If the cost of living went down concurrently then it wouldn’t be so bad, but while TVs and cell phones and computers and clothing have all gotten relatively less expensive over time, housing, healthcare, education, transportation, energy, and even food have not. You know, the things that make up 90% of most people’s expenses.

In in idealized capitalistic economy there is no profit. Profit means there’s an imbalance between producers and consumers, so new companies should enter the market to take up that slack. Of course the companies raking in big profits have every incentive (and the power) to put barriers in front of new competitors, indicating a further distortion of the market. So yeah, profit actually can be viewed as a bad thing.

And employers should care about providing better jobs, because that means more potential customers for their products. Henry Ford knew this, and that was exemplified in the 1950s, which was a period of high taxes on the wealthy, heavy unionization, buying American, and great overall prosperity (at least for white heterosexual Protestant males anyway). That’s the “rising tide that raises all ships”, not new cash being injected directly into hedge fund brokers’ wallets like heroin.

33

But it’s also possible, though not inevitable, that the technology improvements will reduce the cost associated with our consumption categories to more than offset the downward pressure in wages from automation. Fundamentally, this becomes a problem of wealth distribution. I anticipate that we will be able to make more goods and services at lower real cost going forward. Who captures the benefit of that anticipated increase in output? This is an IMHO policy question more than an FQ question.

For what it’s worth, I think we are likely to see marginally better standards of living in the next 50 years than we saw in the last. If recent past history is a guide, everyday people will capture some of the benefit of improved technology even as the wealthy capture outsized portions of those benefits, leading to greater income and wealth disparities. If we see big shifts in political power around unequal wealth distribution, we might see tax and wealth redistribution policies that improve the average person’s life. But I’m not betting on that.

34

That’s what a UBI is generally intended to do, when funded by increased upper-bracket tax rates anyway.

35

That’s absurd. Profit is the whole point of capitalist/free market systems. Why even bother, if you can’t bring in more revenue than your expenses?

And the 1950s weren’t as economically good as they were because of the policies in place. Rather, it’s in spite of them, in large part, that the US was so economically successful. The real answer as to why the US was so successful is because Europe, Japan, and the Soviet Union were in ruins, Britain was having hard economic times, and the US was untouched by war. We were in most ways, the ONLY game in town for manufactured goods and a lot of other raw materials.

I agree that having a culture of good jobs in a nation or an area is something that pays dividends (so to speak) down the line, but it’s not something that a single company can point at their bottom line and connect performance there to the better jobs they’re providing.

I’m not saying that companies shouldn’t provide better jobs, but I’m saying that it’s not self-evident from the metrics they use to measure performance in most cases, except for things like employee turnover rate, productivity and satisfaction.

36

That doesn’t mean that profit isn’t an inefficiency in the market. If you are bringing in more revenue than your expenses, that means you could lower your prices or pay your employees more. Instead, you are pocketting that difference.

In a true free market, this inefficiency of profit is comparatively small. It may be worth your time to run a business, but you’re not getting rich off of it. If you are getting rich off of it, then others will be signalled to move into this lucrative industry and undercut your prices or poach your workers.

We do not have a free market in most industries, and there are substantial barriers to entry. This allows some companies to post absurdly high profit margins, as they rely on monopoly, regulation, and other barriers to entry to keep others out.

I don’t agree that it was in spite of the policies in place. Yes, other factors were important as well, but having a populace with disposable income, rather than wealthy aristocrats taking it all for themselves, helped to grow the economy substantially.

Sure, you want to pay your employees as little as possible, but you want others to pay their employees enough to afford your prices. This is a situation that can really only be rectified by governments ensuring that the people get a living wage, either through regulations on pay or some sort of UBI system.

37

I would say that defending the system by saying “bosses are in business to make profits” is a good reason to start asking, “if treating workers–people–like things is the way the system is supposed to work, how much do we workers–the vast majority of the population–need this system?”

38

I realize that summarizing a book leads to oversimplification, but this is an alternate world to me.

None of the railroad crashes of the 1870s, the Depression, or the Great Recession were caused by overproduction. Each has a primary cause and that was the money people getting seduced by the huge piles of wealth that seemed at hand and getting greedy. In each case, all norms and every law that was on the books were broken by people deliberately issuing misleading statements to gull investors. (In each case a pile of new laws were enacted to cover the old infractions but couldn’t foresee the future ones.)

There has never been such a thing as a free market. Governments have always stepped in, both to create the most favorable possible conditions to to protect the wealthy and to issue restraining laws once the consequences of those actions became too obtrusive. Scientists make a joke about postulating a spherical cow in order to ease their calculations. Capitalism under any definition except that of a spherical cow is about profit and nothing else. That has had widely successful results and horrifically debasing ones, but it is the system we live under.

Supply and demand are real effects in an economy. The terms are usually applied to production, but can also be used for labor supply and demand. Mechanization has in the past balanced the demand for labor by supplying new jobs. As I said above, that is not a simple equation because one new job is not necessary equal to one old job. What the outcome of automation will be is unknown. But the answer won’t come this interpretation of history.

39

The argument is that such speculation/debt/credit/bubbles came about because the base economy had lost steam. The wealthy had accumulated too big a share of the productivity gains, and their only investment vehicle was risky financial instruments because the “bread and butter” industries seemingly had nowhere to go due to market saturation.

40

Seriously, the argument is that, “nothing legal was returning enough money so we had to turn to illegal speculation, because what else could we do?”

I know, I know. People actually said that. I’ve read my history.

It’s a Trump level excuse. Anybody who uses it should be thrown into the nearest jail and made an example of.

I know, I know. They tried that and it never worked. Greed always wins out over memory.

And the explanation may very weakly have some power to simplify what happened in 2008, but it simply doesn’t hold in the earlier eras any more than overproduction does.