As production in the United States evolves to include more machines, programming, and robots, will that have a negative long-term impact on unemployment? Economist Greg Clark says yes, and recent evidence suggests that it may already be happening.
From Economist Gregory Clark's op-ed in the Washington Post:
No, the economic problems of the future will not be about growth but about something more nettlesome: the ineluctable increase in the number of people with no marketable skills, and technology's role not as the antidote to social conflict, but as its instigator....
For much of the past 200 years, unskilled workers benefited greatly from capitalism. Before the Industrial Revolution, for example, skilled construction workers earned 50 to 100 percent more than unskilled laborers; today, that premium has fallen to 33 percent in the United States. The era of the two world wars, 1914 to 1945, was one of particularly sharp gains for the wages of unskilled workers, relative to the rest. Why have the unskilled fared so well? After all, machines -- whether steam engines, internal combustion engines or electric motors -- have replaced people as deliverers of brute force...
But in more recent decades, when average U.S. incomes roughly doubled, there has been little gain in the real earnings of the unskilled. And, more darkly, computer advances suggest these redoubts of human skill will sooner or later fall to machines. We may have already reached the historical peak in the earning power of low-skilled workers, and may look back on the mid-20th century as the great era of the common man.
A couple of things.
Robot Dystopia
Let's start this off with a quote from one of my favorite movies, the film-essay documentary Los Angeles Plays Itself. It's a movie about cinema and representations of urban spaces that is made entirely of clips of movies set in Los Angeles with a narration over it describing how those movies both represent and inform our visions of Los Angeles and the larger country. During the scenes where the narrator discusses the movie Blade Runner (1982) and the sexy killer human-impersonating robot played by Daryl Hannah, he makes this important point:
Perhaps [Blade Runner] expresses a nostalgia for a dystopian vision of the future that has become outdated. This vision offered some consolation, because it was at least sublime. Now the future looks brighter, hotter and blander. Buffalo will become Miami, and Los Angeles will become Death Valley at least until the rising ocean tides wipe it away. Computers will get faster, and we will get slower. There will be plenty of progress, but few of us will be any better off or happier for it. Robots won't be sexy or dangerous, they'll be dull and efficient and they'll take our jobs.
I like the point that our imagination of a robot dystopia being either sexy-dangerous and/or thrilling protects us from a currently unfolding reality. Perhaps we want to think of robots as shapeshifting liquid-metal assassins who travel through time to kill us while we are still rebellious teenagers, or blonde bombshells in red dresses who want to seduce us to steal the military codes so they can nuke the colonies, precisely because that vision is more comforting than the reality that machines are fairly boring bits of code that will generally replace our jobs, depress our wages, and bleed our 401(k)s in ways that are largely invisible to us.
Structural Shifts
The issue of a future in which there are large parts of the economy that are underemployed, unemployed, or unemployable is a serious issue. And the data already suggests this:
(source) Notice how after the last recession in 2001 the number shifted upwards. The boom year of 2006 have an additional 5% long-term unemployed than the boom years of 1998. If you go back even further in that graph, to the 1960s, you see an even larger structural shifts upwards. Here's University of Chicago Economist Kevin Murphy thinking through this issue.
It's difficult to look at, for example, the very low unemployment rates we saw in the early 2000s and say that represented an economy in which everyone was working. Unemployment rates were at roughly the same level that they were in the late 1960s, but if you look at prime-age males, the fraction actually working who were, say, 30 to 40 years old was quite a bit lower in 2001 because there was a big increase in the number who were out of the labor force in that age category...It was primarily low-skilled workers who had withdrawn from the labor market...the opportunities in the labor market for low-skilled workers had deteriorated quite a bit with the rise in demand for skill and fall in demand for low-skilled workers...What that meant was, from a pure labor market perspective, the unemployment rate really wasn't indicative of what the economy was like. Unemployment in an economic sense wasn't as low as unemployment in a measured sense.
The blogosphere is taking note: Add Kevin Drum and Reihan Salam to people who are also concerned about the growing rate of the perpetually unemployed in our economy. The unskilled are slowly disappearing from the labor force, and it isn't clear why or how this is happening. Is it the result of the computerization of the economy, as Gregory Clark suggests? Is it the result of the criminalization of a large part of the poor through specific projections of state power, particularly the War on Drugs and Broken Windows? Is it a more general trend towards higher skills as well as forms of (disability) insurance as Kevin Murphy argues?
What is actually going on here, and if it will get worse before it (ever) gets better, should be a focus of research for those looking at our labor markets.
Historical Markets
Maybe this will all just be fine. The increased presences of computers and robots in creating automation and decreasing the need for labor will increase the need for new types of labor. In particular, the labor necessary for the creation, maintenance, and deployment of said machines. But this work is highly skilled, and there is good reason to believe that it will employ fewer people post machines than before.
Transformation of the skill level of that much labor is going to have large destabilizing effects. The last time we had such a transition of the skilling of labor we'd have to look at the Industrial Revolution in Europe. There we had a transformation from an artisan class into a deskilled, more productive class. This, from an economists point-of-view (but certainly not everyone's), was a huge win for unskilled labor. This new change in the skilling of labor is not what is happening now. Now the economy is taking the labor of an unskilled class and replacing it with machines. In order for someone to take advantage of the new employment opportunities opening up it will requires a peculiar new type of skilling - the ability to create and maintain the new efficiency machines.
It isn't clear what happens here. We could assume that markets will hand-wave these problems away, that wages and incentives and laws will find everyone optimal work, or we can learn from History that these radical changes have massive changes in our lives and begin to prepare accordingly. I'm rooting that we go the second route.
SF writer Jerry Pournelle (at jerrypournelle.com) has been worrying about this for years, as have I. I've worked in industrial automation.
i too have been concerned by this issue for many years. our elected leaders and titans of capital trumpet education, education, education! but we know that this is nonsense; there are not enough skilled jobs available for everyone, and not everyone has the innate ability or desire to take a skilled position. ultimately it's just another facet of the most pressing existential problem of our generation: too many people, and not enough jobs. historically, the solution was war. not shooting up a helpless third world country, but real fire and brimstone war between major powers. the twentieth century is the best example of this; the world enjoyed decades of prosperity in the wake of the slaughter and mass destruction of world war II. but the development and proliferation of nuclear weapons has rendered impossible this method of culling the population. what shall we do? when this recession first started out, i predicted that much of the growing unemployment would be permanent; that the united states would see a new and much higher baseline for structural unemployment. this article highlights just one of the many reasons why.
STC / wiredog: You bring up good points, however your statement makes one common, but foolish assumption. That every job ever created has already been created. As a company replaces its unskilled labor for machines because the labor is too costly, the savings is the put into research for new products and lower prices for their goods/services. The lower prices allow consumers to buy more products (hence the need for research / design for new ones). While this process takes time and it does nothing for unskilled laborers who are now out of work and not qualified for the new jobs, it does help society as a whole. Pittsburgh is a great example of this. It went from unskilled steel manufacturing to high tech biomedical research center and is now one of the cities best handeling the recession. It took time and was very painful (it is still ongoing), but it was certainly worth it.
And a lower population does not increase economic stability. In fact it has the opposite effect. Japan and Russia are both in panic because they have aging populations and once they retire, they know their economies are going to have problems as productivity decreases and there are less consumers for locally produced products.
While this process takes time and it does nothing for unskilled laborers who are now out of work and not qualified for the new jobs,
That's where I (and Pournelle) see the problem. What are we going to do with a large surplus of unskilled labor? I'm talking about the portion of the population that topper out, intellectually, in high school. May not have completed high school. What do we do with them? Welfare Islands? Soylent Green?
i'd really like to believe it, RSS, but it sounds like more of the same hopes and dreams i hear from my boomer parents and extended family who spent too much time growing up reading the utopian science fiction. when have you ever seen a company drop prices due to gains in efficiency? the savings that companies obtain from laying off unskilled labor mostly go to profit-taking by executives and stockholders. research and development is rapidly going offshore, or is done increasingly by disposable H1Bs here. the few jobs that are created by automation in the workplace don't even begin to cover for the thousands of unskilled jobs that were destroyed. none of what you are saying gets to the crux of the issue that 10 jobs created for phd's isn't worth anything to thousands of folks who used to work in steel mills there around philly.
and increasingly, this isn't just about jobs that are considered by most to be skullduggery, or only about folks who dropped out of high school, maybe just scraped by to get their diploma. between automation, offshoring, and outsourcing, college grads are getting nailed too; exactly the sorts of knowledge workers that our economy was supposed to have needed. and unlike the high school grads, a lot of these college graduates racked up a lot of debt to get their credentials.
for the last thirty years, it's practically been an unstated policy of the american government to aid and abet offshoring and the stagnation of working and middle class wages. if we're gonna fix this outside of simple supply and demand on the labor market - lowering the population - it's up to the american people and the american government to consciously change their policy to one that emphasizes job creation over raw profit, protectionism over placating our trading "partners", and the dignity of labor in general.
sadly, i'm not gonna hold my breath.
It is very much a structural issue, and its been recognized in its general form back to Karl Marx, more specific to America by economist Kelso (with Adler, The Capitalist Manifesto). Undeniably, management will always try to minimize labor costs. As technology enables that, more become unemployed. Certainly training for the new technologies helps, but advanced technologies may require skills beyond the abilities of many job-seekers. You can add the offshore trend in software and support as, again, labor costs are minimized. //
What to do? Well, the problem isn't the technology. Being more efficient, doing things for less, shouldn't be a problem - that's what competition is about, right? The problem isn't robots (or whatever technology) taking jobs, it's the need for jobs in the first place. Whoa! Socialist? No - capitalist! There are a number of people in the US that don't have jobs and don't need them-- they are capitalists. They own income producing capital (companies, which own factories, robots, etc.) which produce the value-add they live off. So the problem is the ownership of the robots/technology. If every displaced worker got a slice of ownership of that robot/whatever that just replaced them, and a dividend check from it, then they may not need a job so much. The private business way to do that is well established: Employee Stock Ownership Plans (ESOP's), where a loan to the employees is paid back from the dividends/profits. Its not a give-away, its a loan.//
So rather than create a class of chronically unemployable citizens who need welfare and don't have money to consume what the economy produces, a citizen's ESOP, a Capital Homestead Act (like what Lincoln's Homestead Acts did for land), would allow Joe Public access to credit that the rich enjoy, and in a structured manner, enable us to participate in our capitalism, with real capital, real value-add, and then, the unemployment rate wouldn't matter. We could have our technology and our income too. For more information, please see www.cesj.org, or Google "Capital Homestead".
How about if there are eventually only 100 people (or robots) earning money and each earns $10 trillion per year, taxed at 50% and the tax money is delivered to each of the other 10 billion people to top up incomes up to $100,000 each. Each person does whatever they want with their time and money, and will know they will earn at least $100,000 even if their projects do not make any money.
I'd also counsel against assuming that government "will hand-wave these problems away, that wages and incentives and retraining and laws will find everyone optimal work." Laws don't work, retraining programs don't work.
Lump of labor fallacy.
Your thinking is muddled here. Technology, and the ability to produce more with fewer workers and more machines, is precisely what has led to the great increase in standard of living. In the long run, wages are productivity. You mention the Industrial Revolution, where machines meant that the unskilled could suddenly produce more than previously the very skilled could do. But even since then, there has been increasing automation and fewer workers operating more machines for higher wages, while the work has remained unskilled. When we say "unskilled," these jobs often require skills and knowledge, but the sort that can be taught to nearly anyone, and taught on the job and fairly quickly.
The problems that Gregory Clark point to are twofold:
First, while most technological developments, even in the 20th century, were about increasing the productivity of unskilled labor, are recent developments available only to the skilled (or the intelligent, or otherwise to a limited subset of people)? If so, then the productivity of those who can operate the machines will soar above those who cannot, increasing inequality. But would we still choose inequality anyway, if the productivity gains were so high that we could still make everyone better off than if we didn't have the new technology?
Second, is the pace of technological change too rapid? The Industrial Revolution was wonderful for standard of living in hindsight, but to be caught up in it was not always great for people who had spent their lives learning to be artisans when the world shifted on its axis. Are we undergoing revolution after revolution, so that we never have a time to sit and enjoy the fruits before there's another upheaval, where previous skills are rendered useless?
Combine the two, and you can indeed result in a worse problem-- what if the skills necessary to get maximum productivity take longer and longer to teach, but become obsolete as quickly as they can be taught?
I think Rand W is on the right track here - The parallels between nature and economics are a cliche at this point, but the economic solutions that pay the most surprising dividends still tend to as elegant and strange as the paths taken by evolution. Unfortunately, humans being the way they are, many people tend to balk at unusual solutions and expect a governmental Rube Goldberg contraption on a scale equal to the problem, and aren't satisfied with merely tipping over a specific, well-hidden domino. A pleasing irony might in fact be using supercomputer simulations of the economy to find that domino and save us all from obsolescence.
We've already passed the point of no return, so I think our best bet is not so much how to minimize the damage using existing methods as it is to spin these new economic developments in ways that are advantageous to as many people as possible. Of course, that's just a platitude, but I think embracing automation and selling it as a new kind of decadence would appeal to both the "conspicuous consumption" and "aspiration" crowds.
After all, reading the commentary of the Dead White Men who set us on this track in the first place makes it pretty clear they were looking for release from the metaphorical Curse of Adam, and centuries later the '50s were overwhelmed with visions of the Future-with-a-capital-F. The memory of society is too short to stave off disappointment if promised developments don't happen within a decade or so, but enough time has passed now (the article even mentions being done with traditional dystopias) that I think people won't find a kind of mild utopianism unpalatable - Something to focus people's desires and aspirations on the long term, so that they'll still want to have a stake in society and not become disaffected or criminal.
I love the idea of a basic guaranteed income (especially as a somewhat struggling artist...), but in the foreseeable future I don't think it's politically tractable in the US. What if, over the next decade or so, there were incentives for companies to develop and market technologies to help the unemployable live somewhat "off the grid," like solar panels and 3D printers, or even small automated greenhouses (I've been involved with projects like this, and here's a link: http://mama.agr.okayama-u.ac.jp/lase/tomato.html )? After the initial buy-in, which could be subsidized as something Green, the costs of maintaining and operating these technologies would be a steady, affordable trickle that wouldn't overly tax the resources of the poor but would still keep them invested in society, especially if they see the rich dabbling in more elaborate home automation that they might one day come to possess. ..And while you could argue that using technology to make people too independent of the economy would bring about further financial ruin, it seems to me that the very ambition that drove these developments in the first place would keep people reinvesting in the economy and developing further new technologies and jobs, until, I suppose, the world becomes too Sci-Fi to talk about clearly...
Good lord, I'm proofreading this and it all sounds so cynical, but there are no wrong answers in brainstorming, I hope...
Robotic technology is going to make mincemeat out of all extremes of ideologies (socialist, capitalist, etc), since pragmatism of how the world functions is going to overtake ignorant idealism of the idealogues.
Josesph Schumpeter was dead on when he analyzed economic trends, I imagine him analyzing trends in robotics would only cement his views today that capitalism as we know it today is slowly going to lose it's relevance
"Schumpeter's theory is that the success of capitalism will lead to a form of corporatism and a fostering of values hostile to capitalism, especially among intellectuals. The intellectual and social climate needed to allow entrepreneurship to thrive will not exist in advanced capitalism; it will be replaced by socialism in some form. There will not be a revolution, but merely a trend in parliaments to elect social democratic parties of one stripe or another. He argued that capitalism's collapse from within will come about as democratic majorities vote for the creation of a welfare state and place restrictions upon entrepreneurship that will burden and destroy the capitalist structure. Schumpeter emphasizes throughout this book that he is analyzing trends, not engaging in political advocacy."
"One of the great advantages of capitalism, he argues, is that as compared with pre-capitalist periods, when education was a privilege of the few, more and more people acquire (higher) education. The availability of fulfilling work is however limited and this, coupled with the experience of unemployment, produces discontent. The intellectual class is then able to organise protest and develop critical ideas.
"In Schumpeter's view, socialism will ensure that the production of goods and services is directed towards meeting the authentic needs of people and will overcome some innate tendencies of capitalism such as conjecture fluctuation, unemployment and waning acceptance of the system"
http://en.wikipedia.org/wiki/Schumpeter
For the past hundred years, every wave of automation has been greeted by exactly the same arguments we see here. Critics said the automobile, the washing machine, and the personal computer would "take away all the jobs".
I remember a high school professor lecturing that half the people in the US would be out of work by the year 2000. He said he had the undeniability of statistics on his side. When I first graduated college, I trained secretaries in the use of the Apple Macintosh as as a word processor. In every single case, without exception, they were convinced that the computer was going to put them out of work.
Yet in all those 100 years, none of these predictions, the high school professor's statistics, or the worries of clerical workers have EVER come true. In fact, our economy has for the past 100 years been a job creation dynamo and employment has been growing exactly fast enough to fit the growing population...until recent economic problems, which are clearly based on outsourcing and financial trouble, not technology.
Yet here we are, going over the same ground again. Again people are arguing, despite 100 years of proof to the contrary, that automated tools (this time, robots) are "going to take away all the jobs".
Sigh.
Folks, lets get back to reality, not rhetoric. If we look at the history of the world, without exception, the societies with the largest amount of automation have the MOST jobs and the HIGHEST pay. The societies with the least amount of automation have the LOWEST number of jobs and the LOWEST pay. If you want to see societies with no robots, go to Haiti, Mexico, or Nigeria.
So let's start this debate over.
Let's step back and examine the "robots are going to cause permanent unemployment" arugment, and try to understand the fallacious assumptions that lurk behind it.
Why do so many people believe something that is so obviously, from experience, not true? That is a worthwhile question.
"Let's step back and examine the "robots are going to cause permanent unemployment" arugment, and try to understand the fallacious assumptions that lurk behind it."
Except there aren't any fallacious assumptions that lurk behind it, instead there is a mounting body of evidence of permanent unemployment. You're taking a snapshot from a different era and trying to apply it to trends we have much foreknowledge about: Eventuality of AI and robots that can do menial labor much faster (and therefore cheaper) then most humans WITH HUMAN capable intelligence.
Yesterdays automation lacked sophistication of a machines ability to be intelligent, they were "dumb machines". Next generation automation is going to full autonomous and that is going to have sever consequences for human beings, since fully intelligent machines are now capable of out competing human beings economically.
We may not see it in our lifetime, but our children or grandchildren will definitely have to deal with the issue eventually.
This is coming to intellectual labor soon, you're missing the fact that when we've automated, even "lower skilled jobs" still require some amount of intellectual labor, once machines surpass human intellect the vast majority of humanity will no longer be economically viable.
Robots do not have biological drawbacks like getting tired, feeling pain, and whatnot. Businesses will have an intelligent workforce with a body that does not carry biological deficits.
Sooner or later artificial intelligence is going to make all but the most intelligent and wealthy "unskilled labor", I really don't think you can even grasp what has occured, nor what is going to happen.
Yes, I understand your argument. All these arguments were made in those earlier eras. Everyone was so sure we would have flying cars, and that congestion in the skies would be the scourge of the future. But guess what...we didn't have flying cars. We probably never will. The people who made those predictions were assuming a linear growth rate of technology that in fact was not realistic. They didn't understand some of the very real underlying limitations on flying, which we now know prevent the common use of flying cars.
In a similar way, there are many who are assuming robotic function which simply does not, and will not soon, exist.
Is it "inevitable" that we will have "human levels" of intelligence? Absolutely not. That is pure myth, promoted mostly by people who aren't practicing AI researchers. Contrary to all the breathless predictions, in fact we are as far from true machine intelligence as a primitive stone age tribesman would be from building a 747. We don't even know what we don't know. Moores law does NOT drive AI. Increasing processor speed and evolutionary algorithms do NOT cause intelligence to simply pop into existence. It may well take a hundred years and the entire world's technological base to accomplish even a primitive artificially intelligent computer. It is that dificult a challgenge and no reason to assume that shortcuts exist.
I and the other real practicing roboticists of the world can only sigh and shake our heads, and hope that some day the rest of the people in the world revisit reality.
Meanwhile, I will predict this. The societies that continue to embrace advanced automation and technology will continue to see huge job creation, growth of income, and growth of social equity.
There will be economic problems, class struggle, arguments over taxation and the rich vs. the poor. But those issues are politicical, not technological. If one believes in socialism, fine, pursue socialism. That is a political and economic decision. But don't waste time trying to drag robotics into it, because it's unrelated.
"Yes, I understand your argument. All these arguments were made in those earlier eras. "
All of those earlier era's did not have even a fraction of our computing power which is doubling roughly every two years. As for flight, the real problem with flying cars was a misunderstanding of it's cost and of nature and the fact that chemical engineering is in the stone age, flying cars could become quite easy to make with necessary advances in chemical/materials engineering that have not yet taken place, one has to understand the scientific state of a science and it's underpinnings to really grasp the possibilities before one can even think of making projections.
You're also missing the fact that people in the past did not adequately understand technological trends by looking at the sciences behind it, as we go back in time those projections were based ignorance, not sound knowledge.
We already have AI in the form of countless forms of life from bacteria, to small animals to human being. Once the brain is reverse engineered sufficiently enough we will soon have have greater then human intelligence AI on our hands, even if we are only simulating a subset of the brains power it matters not because that subset is still going to displace whoever was doing that work previously.
Your argument is not like the flying cars argument, your argument is more you arguing against the limits of transistor process technology. There were always naysayers at every x stage in moving to new transistor shrinking technology. AI does not have the limitations flying cars do with regards to the state of the science of chemistry and materials sciences (in regards to flying cars).