The Myth of the Myth of Technological Unemployment

by Andrew McAfee on January 23, 2013

Over at Slate, Matt Yglesias has a post titled ‘The Myth of Technological Unemployment‘ accompanied by a graph showing that hours worked in the US have been rising and falling in lockstep with output. He writes

Machines are replacing workers, in other words, but they’ve been doing so since the cotton gin and the spinning jenny.

Which is absolutely true and completely uncontroversial. It’s also true that previous waves of automation have not, in the long run, led to mass unemployment. But is that still true? Here’s a graph (drawn with the assistance of my trusty assistant FRED) similar to Yglesias’s, but concentrating on US manufacturing output and jobs over the past 40 years.

FRED Graph

That really looks like technological unemployment to me, especially when manufacturing employment is also on the decline in Germany and Japan, in China, and around the world. When this is the case, it means that employment changes are not due to jobs moving around the world in search of cheap human labor; they’re due to machine labor becoming at least as capable as and cheaper than humans.

There’s evidence that the same phenomenon – rising output but falling employment – is also occurring the US wireless industry, at least in part because of “gains in productivity and the fact that more customers were going online to choose their phone and pay their bills, reducing the need for more call-center workers and salespeople.” Again, that looks like technological unemployment.

How’s the US economy as a whole doing? Take it away, FRED:

FRED Graph

We are still adding jobs and working more hours in non-recession years, but not as quickly as we used to. Since the end of the 2001 recession real GDP has increased by just about 20%. The number of hours worked, however, has increased by only 2.8% over that same time, and the total number of jobs by 1.9%. Those latter two numbers are pretty close to zero. Is it so hard to believe that a realistic future combination of fast automation and relatively slow GDP growth could cause them to turn negative?

I don’t find that scenario implausible at all. Technological employment has not happened economy-wide yet, but as the facts change — as technology’s role in the economy shifts — shouldn’t we change our opinions about what constitutes a myth?

Michael Lorton January 23, 2013 at 11:56 pm

Is it so hard to believe that a realistic future combination of fast automation and relatively slow GDP growth could cause them to turn negative?

You mean, if we assume automation causes unemployment, is it credible that automation might cause unemployment in the future? Yes.

However, if you don’t assume the thing you’re trying to demonstrate, then it becomes more difficult to believe that something that has never, ever happened before, despite the constant warnings, will suddenly start to happen.

Abdella Ali January 24, 2013 at 2:33 pm

Well I think it’s also a little short sighted to consider ‘technological unemployment’ as improbable simply because it has not happened yet.

I wouldn’t even begin to make predictions like McAfee, I don’t have the mental acuity to consider all the factors and variables required to do so, at least I don’t think. However I do at least think it is important to consider that there are fundamental differences between the technological booms of the past and the technological booms of today – things like the time it takes for a technology to go from R&D to obsoleting a particular job. Or the fact that an ever increasing amount of employment involves or even relies on computer use, which to me seems to open it up to software automation with fewer obstacles.

Fundamentally, I think that it’s not just this idea that we are getting ‘better’ at making technology do things that humans do (we are), but that it’s happening faster and faster, maybe faster than we can reasonably be expected to keep up. That it’s happening with a reduced barrier of entry; a fresh-faced software startup that creates a product can fundamentally change the dynamics of an office isn’t some crazy concept anymore.

So… maybe it would be prudent to be a bit less critical of people like Andrew’s ideas.

George Zapo January 28, 2013 at 1:46 am

After viewing the graph, numbers, and relative facts in this article, I have to agree that technology, machines, and automation are a primary reason more people are unemployed, and may stay

I question what might happen with so many people out of work around the world.

Anonymous February 3, 2013 at 10:57 am

The main reason automation is adopted is because it is less expensive than labor. Many industries didn’t automate, but moved jobs to Asia where labor rates are less than 10% of US labor costs. Perhaps we should eliminate the minimum wage to drive higher employment in the US.
If so, how would people in the US make a living on $1/hr? I realize that’s not possible. I’m not a fan of welfare, but we might be better off using the EITC to support people who are being productive, rather than paying people to not work.
Lot’s of difficult choices – I look forward to following this continuing discussion.

Cirith February 5, 2013 at 8:14 pm

And china has also started to invest in more automatisation. Well I think that many who still think technological unemployment is impossible just have no idea what technology is capable of. Hell 3D printing for houses isn’t quite ready yet, but it’s already clear that it will come. And while hard AI is still far away, there is more and more we can do with soft AI. The difference to the past, is that in the past most technology only amplified human productivity, with the right soft and hardware humans can become almost unnecessary in many fields.

And quite honestly when we can do more and more with less humans, why would we wan’t to still work the same amount? If say a thousand people could produce all the goods we have today for one million people, would it really make sense to try to find enough new things for the others to do and make everyone keep working 40 hours a week? Time is a scarce ressource, at some point we should use our productivity gains to give everyone more free time. Afterall we aren’t producing for productions sake, we are making things for humans to use.

Anyway I believe that most work can be automated without hard AI and that is a great thing(not that I think hard AI isn’t possible, but it’s very hard to predict when we wil reach thet point, maybe there will be a breakthrough in 20 years, maybe in 100 or maybe in 10). We just need to adapt, else the transition will be ugly.

Matt Beane February 10, 2013 at 10:11 pm

Erik Brynjolfsson’s (and others’) work on GDP seems to indicate that it used to be an acceptably lousy proxy for general consumer welfare, and that recent socio-technical changes (e.g. the internet) make it an increasingly-unacceptable one. How big a factor might this “hidden welfare” be? Big enough to dampen – let alone counteract – the trends in your FRED graphs, Andy? My impression is no…

Warren Harding February 11, 2013 at 7:34 am


Our governments should be spending more of our tax-dollars on something many seem to want. I’m for robotics that are owned by all of the citizens of a country. Lot’s of people want to cheer every time a robot puts someone out of work. Fully automated robotics factories, with self replicating robotic arms. Highly automated renewable energy, windmills or underwater water mills. Highly automated steel production. Highly automated chip manufacturing, and Linux. I’ve seen some automated building manufacturing companies starting up as well. Other prerequisite products can eventually be manufactured as well. All source code and blueprints have to be fully owned with rights to an infinite amount of use. All owned by the citizens of the country concerned. Small factories at first, with all of the bugs worked out, so that it largely builds itself in the end. It should be affordable, I’m an economic conservative. Eventually the complex can produce consumer goods besides steel, energy, chips, buildings, and robotics. Charities and the open source community can help as well. I support liberal licensing agreements of source code and blueprints, to allow royalty free replication.

Anonymous February 17, 2013 at 3:18 pm

However, if you don’t assume the thing you’re trying to demonstrate, then it becomes more difficult to believe that something that has never, ever happened before, despite the constant warnings, will suddenly start to happen.

You mean, like global warming?

Besides, it has happened before, it’s just that before there were other places to go. When agriculture was mechanized, manufacturing stood ready to absorb the unemployment to make the machines that mechanized agriculture. But then what? What happens when not just one thing (agriculture) is mechanized but everything else is too? Where do you go then?

Deniers of technological unemployment blandly state that “something will come up.” But what? Where? When? When agriculture started to decline as a source of mass employment, there was no waiting for the next big thing. Manufacturing was there already.

Curt Welch February 24, 2013 at 7:29 pm

We should not be calling this Technological Unemployment because that misdirects people’s ideas to the wrong place. The economy will always be able to create jobs for people (though we might have to do away with minimum wage for that to work). That’s not the problem.

The real problem is Technological Inequality. The advances in technology are not being shared. They are driving increasing levels of inequality. We must stop focusing on the idea that jobs are being lost, and focus instead on the true problem – that our income distribution is being polarized with a small minority of people able to capitalize on the new technologies and drive their wealth and income to astronomical levels, while the bulk of society, only sees their wealth and income fade towards poverty as their work is displaced and downgraded by increasingly complex machines.

Factory workers making $50k that get displaced by machines can still find work. They just have to get a teaching job for $15 an hour, or work at Starbucks for $8 an hour. A few factory works might have the drive and brains to start a new business and make more than they were making in the factory, but most, will be forced in the other direction to lower income work.

Automation and technology is hollowing out the middle class, with only a few finding a path to the top, with most falling towards the bottom. This trend has been going on for 100’s of years – throughout the entire industrial revolution. We have been offsetting the problem with increasingly complex and large social programs and laws. Antitrust laws, labor laws, legalized labor unions, progressive taxes, minimum wage laws, free education, child lunch programs. The list is huge. Everyone one of these programs offset the problem by taking more from the top, and sharing it with everyone, which means the rich pay more and the poor benefits from the money taken from the rich. It’s just wealth sharing in the form of government programs.

But around the 1070’s the rich put their foot down and said enough is enough and talked the US people into supporting a change in direction by reducing the burden of the rich by reducing their taxes, reducing regulations, and by reducing social programs that benefited the poor. And the result has been a disaster which led to such high inequality that it allowed the rich to get so greedy, they took down the world’s economy – and then they had the nerve to say the poor were responsible for paying for the damage. And in a big way, the rich are correct, the poor are responsible because they are the ones that voted to give the rich all their money in the first place.

Technological Inequality is not a short term blip in the economy. It’s a game changer that will never go away and never get better. Technology, once invented, doesn’t go away. It will get worse and worse and worse – and it will happen with increasing speed and disruption every year as we move forward. Technology advances exponentially so the problem not only gets worse every year, it gets worse faster, than the year before.

Either we let the wealth of our nation, as well as the control of our government, fall only into the hands of the few, and let everyone else struggle with a standard of living heading towards poverty, or we fix it by turning back the bad policies of the last 30 years, and get back on track with constantly increasing the amount of wealth distribution we do in society.

The inequality is so bad, that it really can’t be fixed with things like minimum wage, and progressive taxes anymore. No amount of tax relief at the low end fixes the problem of someone making only $8 an hour and only working 20 hours a week and receiving no health care or retirement benefits. Removing all their tax burden will not given them a living wage. Raising minimum wage is an approach, but for many, it will just force the company to force cut backs of works hours and job slots instead of creating increases in take home pay. We also don’t need yet more government programs for the poor. All the government programs tend to just be wasteful and inefficient – proving services they don’t need to some, while others in desperate need of help get none. While at the same time, having unfortunate side effects of locking welfare recipients into a welfare live because of negative incentives to work created by removing their welfare when they start to work.

The correct way to do redistribution is to drop all these programs, and just redistribute cash directly to EVERYONE in the country as a Basic Income distribution. Not to just the needy, but to everyone. It’s a welfare for everyone, that never goes away. It’s not conditioned on being unemployed, so it doesn’t go away if you get a job. If you work, your wages are just on top of of the BI payments. If you get fired, you lose your wages, but your BI is still there. No need for unemployment insurance. People just need to take the risk of getting fired on their own. It’s socialism, because if forces the rich to help out everyone in the country. But it’s fairly different from typical welfare because the money we pay in taxes goes to everyone – our children, our wives, our friends, not just some “lazy guy in the slums”.

The level of wealth redistribution we already do in the US, is so large when you add it all up, that we could convert most the current programs to a citizen wide BI distribution, and end up giving everyone in the country, a payment of around $1000 a month, or $12K year. We could do that without raising taxes, but by instead, only converting over all the current welfare programs to a BI program. Social security should be replaced with a BI like this so everyone gets the same fixed income from the government no matter how much they were able to continue over the years. Minimum wage should be elimited allowing businesses to hire people for any amount the market allows helping to drive unemployment towards zero.

With the BI payment, we would no longer need progressive income taxes. We could switch to a flat rate tax that allows every dollar we make, to be taxed at the same rate. The flat rate tax, combined with the BI, crated an automatic progressive redistribution effect so we don’t need progressive increasing tax structures which greatly simplifies tax ramifications for businesses. All personal income can be taxed at a flat rate, whether it be wages, salary, capital gains, or distributions. With a BI to all family members, most complex tax deductions meant to help families are no longer needed. If you are single, you only get on BI payment. If you are a family of 4, with two grand parents living with you, the family is receiving 6 BI payments to help support the entire family. No need to give complex tax breaks to families filing jointly.

Yearly income taxes filing can be elimited due to the fact that all income is a flat tax. You are just required to pay the tax at the time the income is received as payroll deductions and there’s no year end calculations to be done. Most of the IRS can be eliminated.

When everyone kin the country is receiving a basic income payment, people in poor areas can help each other. They can move in together and share the cost of housing and food even if they are having problems finding work in the area. When everyone has some basic level of income to spend, there will be no areas of total poverty. Even if jobs are scarce the poor areas will still need stores selling food and basic living goods and that will create jobs in the poor areas, which will lift the standard of living of the area.

With a BI payment, people will be mobile If they get laid off, and want to move to a different city to find work, they know they will have money to pay for a room to live in, and food to eat while they search for work. Or at least, be able to pay for gas, and food, while they live in their car. Today, if you get laid off, and have no income, you are forced to live off the good will of friends – and if you have no friends in a distant city where there are jobs, you have no way to bootstrap a new life elsewhere.

A BI is not just free money for the lazy. It’s a leg up, we are willing to give to everyone so that the best and brightest of our society, will have an opportunity to succeed, no matter how unfortunate their life happens to be. It will allow us to fund moms, so they can opt to stay at home and raise happy healthy, productive kids, instead of having single moms struggling to support children they get to spend little time with, due to trying to hold down 2 jobs.

No one receiving the BI need to feel like they are leaching off of society because everyone in the society gets the exact same welfare treatment. NO one needs to feel cheated, by not being given a fair “chance” in society. Everyone feels equal in terms of how the society is helping them succeed.

We don’t need to care which of our citizens advances to become the next 100 billionaire because however it is, they will have had to pay the same taxes as the rest of us, and they will have done more good to help everyone, than the rest of us.

A flat tax funding a basic income payment for everyone is a far better way to do welfare than all the complex programs we currently have in place. It’s the solution that can fix technological inequality no matter how bad it gets – because we just keep inching the tax rates up and raising the BI payments any time inequality grows too larger. When machines are doing 99% of all the work, then 99% of the people don’t even need to be working. And the BI is how they can continue to live, when this happens. The people who create the best machines and products and businesses will still get insanely rich, but no one will suffer in poverty at the bottom and everyone will have a safe and stable life. How big a lifestyle above everyone else we manage to live will be limited only by how successful we will be in helping to create our automated economy. There will still be inequality, and there will still be great riches waiting for anyone that devotes their life to making money and who manage to succeed at it, but no one at the bottom of the ladder will suffer when there’s such huge wealth being created by the machines. And anyone, and everyone, that wants to work, will have lots of jobs to pick from, even if it might be little more than an internship because of how little it pays. But emotionally it still gives people the satisfaction of being able to be part of society, and to help others. There’s nothing more depressing, and degrading than not being able to find a job – that is what drives people to be violent sociopaths that act out against society instead of feeling connected to, and a part of society.

Matthew King March 13, 2013 at 10:05 am

Technological unemployment is the goal. With it comes the time to pursue our true passions and interests, spend time with family and enjoy the wonders of life outside manufacturing plants and office cubicles. The challenge is how to distribute wealth in order to realize this dream.

Guest March 13, 2013 at 10:10 am

Tiago Fernandes March 13, 2013 at 10:10 am

Agree, that would be a rational and peacefull transition to a greater good – a Resource Based Economy.

Stefan Sargent March 13, 2013 at 12:52 pm

We actually have been at risk of technological unemployment for at least the past 30 years. There are factors which have forestalled it.

1) Advertising, creating desires and perceived “needs” where none existed before.

2) Extension of credit in the form of credit cards. It actually took quite a sales pitch to get a spend thrift American culture to accept credit cards, but it has allowed the economy to continue producing and selling to consumers, despite a stagnation in wages, and an increase in corporate productivity and profits.

3) Expansion, not only into service sectors, but growth into new markets globally, ie. “development”.

4) The ever controversial “planned obsolescence” .

As you are no doubt aware, it’s not just manufacturing any more, but service, construction, and data mining which are endangered by recent developments in robotics and computing.

The financial economy grows, creates new sectors, and hence employs, only by increasing incorporation of natural resources and human relationships into markets. It seems to me highly optimistic to believe there are no ecological or psychological limits to this incorporation, and hence expansion.

Even if we assume we can perpetually “create new jobs” by opening new sectors from the very technology that is displacing human participation, the geometric rate of improvement of technological efficiency means that it will continue to progress faster and faster. This requires people to learn new skills faster and faster. Any secondary education program I know of takes at least 2 and upwards of >10 years to complete. Already Universities and Colleges are finding it difficult to structure their programs around employment, because by the time students (young or old) finish their programs, the world has changed drastically from the world which existed when they began.

Toby Russell March 13, 2013 at 3:02 pm

And don’t forget stagnant to falling wages. There’s a reason labor is worth less today than 30 to 40 years ago; it’s not demanded as much as it used to be, thus it has less bargaining power. It can only have less power if it is less in demand. It can only be less in demand if what it did can now be done by machines/computers. We certainly don’t have any productivity problems. There are close to 19 million empty housing units in the US alone.

And there’s another huge element to this people hardly ever consider. Do we really want to sustain consumerism and perpetual economic growth (I’m ignoring the planet’s carrying capacity for the sake of brevity) forever, perpetually? What’s so good about economic activity that it and it alone should be blessed with the ability to grow forever and ever (i.e. cancerously)? Economic activity is, crudely speaking, buying and selling stuff. Why is this type of activity So Good that it should expand forever until, logically speaking, it controls literally everything we do for each other (“Pass the salt please.” “Sure, that’ll cost you 2 bucks.”)? If we don’t want this logical outcome of perpetual economic growth (I assert we do not), we should be looking at how technological unemployment can be embraced to (finally!) increase leisure time, and to begin to shrink the economic realm, the realm of buying and selling.

We should be re-imagining money’s role in society as purely an enabler of economic activty (minimizing its function as store of value) and redesigning the money system so that it is comfortable with both economic growth and decline (“steady state economics”), rather than as a systemic forcer of economic growth as it is now (via compound interest and fractional reserve banking). The UK’s “Positive Money” has such a proposal, as does the thinker Charles Eisenstein, the US’s AMI (run by Steven Zarlenga), and there are yet other proposals out there for reforming the money system. They all need to be debated.

With a modernised money system we could fund things like a Basic Income Guarantee, which should be seen as a new, updated way of distributing purchasing power to people (not as ‘something for nothing’), needed if we are to embrace technological unemployment fully, as I believe we should. Today we are far too dependent on wages for distributing purchasing power. This dependence is starting to hurt us as technological unemployment and the end of economic growth begin to bite hard.

The simple truth is that we don’t need each other economically as we used to, but that’s ok. We just need to relearn value and how to be valuable to each other; the monetary measure of
value is less and less helpful in this area (i.e., banking is far more ‘valuable’ in the current money system than parenting or teaching). We need each other as family members, as friends, and members of the community, but wages are no good at nurturing these vital roles. And without sufficient purchasing power, we live a dangerous, hand-to-mouth existence that makes us no use to anyone, a terrible waste of human potential.

We fail to value each other as humans, not because it is impossible to do so, but because our focus is too much on numerical, monetary value (price), which is not subtle enough for
the complex ways we are always going to be useful to each other.

Sorry, this has been a very long comment. If you made it this far, thank you for your time.

Anonymous March 13, 2013 at 9:46 pm

That is why we support a natural law economy called the resource based ecnomy via the Zeitgiest Movement and the Venus Project.

Christian Holata February 19, 2014 at 8:38 am

This ignores that most jobs today are irrelevant and just help to maintain the status quo. We would have an unemployment rate of 75% already but of course you can employ everyone as long as you come up with bullshit jobs. Those bullshit jobs will waste energy, resources, cause traffic jams and depression and whatnot …

If your goal is to work more hours doing meaningless bullshit then you really have no understanding of economics. Look up “to economize” …

Christian Holata February 19, 2014 at 8:38 am

100% ACK

Matt Grant March 4, 2014 at 2:53 pm

Im writing a business paper on technological unemployment and I find this article can be much to my use. Does anyone know what the index is exactly measuring? It would be really helpful to fully understand the graphs.

stand up comedy September 12, 2014 at 5:05 am contoh kata pengantar

Comments on this entry are closed.

Previous post:

Next post: