Everything You Think You Know About the History and Future of Jobs Is Likely Wrong
The polarizing effects of technology on occupational skill requirements since 1979
47 Percent
That’s the highly-cited estimate out of Oxford by Frey and Osbourne of the percentage of existing jobs that are likely to be automated away with the help of technology within the next 15 years. According to this paper, flip a coin and call heads or machines to see if your job will exist in the 2030s. This is the 21st century fear for many called “technological unemployment.”
There is also a conception among many (about half of those considered experts, so again flip a coin) that there is no technological unemployment problem because even though technology eliminates jobs, it also creates new and better ones of sufficient supply such that pretty much everyone is better off than they would be in otherwise “undisrupted” lives.
There does tend to be one caveat to this dismissal of automation fears — that most of all of these high-skill jobs will require high-skill labor, and thus a highly-skilled work force which in turn will require more education. So of course the answer to what could otherwise be a somewhat thorny future, is simply education, education, and more education.
Well, David Autor of MIT published a fascinating paper (though we reach different conclusions) in the Summer 2015 Journal of Economic Perspectives titled, “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” In this paper he compiled the following chart, and it should blow anyone’s mind who holds a strong opinion about the historical effects on jobs by computing technology since it took off in the 1970s, and so the possible future we should expect if the trends continue.
This chart has some very interesting complexity to it, which I’ll attempt to simplify below with snippets from Autor’s paper. The most important thing to recognize is that all points above the flat red line show relative growth in jobs while points below show relative loss in jobs. Points to the left show jobs with less skill required and points to the right show jobs with more skill required. Each curve represents basically a different decade.
Three Important Employment Observations
First, the pace of employment gains in low-wage, manual task-intensive jobs has risen successively across periods, as shown at the left-hand side of the figure.
If we look at the far left of the chart, we see growth in jobs with the least skill, increasing decade after decade after decade. As the story goes, technology should have the opposite effect. The simpler a job is, the easier it should be to automate, and yet we’re not seeing this at all. Instead we’re seeing more and more low-skill jobs being created not destroyed.
Second, the occupations that are losing employment share appear to be increasingly drawn from higher ranks of the occupational distribution. For example, the highest ranked occupation to lose employment share during the 1980s lay at approximately the 45th percentile of the skill distribution. In the final two subperiods, this rank rose still further to above the 75th percentile—suggesting that the locus of displaced middle-skill employment is moving into higher-skilled territories.
Where these curves intersect the red line has been moving to the right, meaning that more and more middle-skill jobs have been lost in a way that even increasingly eats into higher and higher skill ranges. This is a hollowing out of the middle and even upper-middle. Jobs that require what’s considered between a low and high amount of skill have been disappearing. This appears to reflect the loss of the middle class. As each decade passes, the jobs that require mostly a medium amount of skill are simply going away, replaced instead with jobs requiring less skill, not more skill, and thus jobs that tend to pay less, not more.
Third, growth of high-skill, high-wage occupations (those associated with abstract work) decelerated markedly in the 2000s, with no relative growth in the top two deciles of the occupational skill distribution during 1999 through 2007, and only a modest recovery between 2007 and 2012. Stated plainly, the growth of occupational employment across skill levels looks U-shaped earlier in the period, with gains at low-skill and high-skill levels. By the 2000s, the pattern of occupational employment across skill levels began to resemble a downward ramp.
We should expect to see what we see on the left of this chart on the right instead, but we don’t. Between 1979 and 2007, a span of almost 30 years, there was less and less growth in jobs requiring the most skill. Only since 2007 has there been a reversal with a small amount of growth in these jobs. Other than that, as Autor himself describes it, it looks like a “downward ramp”, meaning that both middle and high-skill jobs are being steeply replaced with low-skill jobs, and have been since the 1970s.
This is not the story we are told. Instead we read story after story like this one from the Guardian in 2015, claiming that 140 years of job creation show that jobs will always be created. And yet the ongoing trend of such articles is to mostly ignore the potentially unnecessary nature of the jobs themselves, the level of skill involved to perform them, and the lower pay they can command than the jobs they are replacing. Take for example the following excerpt:
Their conclusion is unremittingly cheerful: rather than destroying jobs, technology has been a “great job-creating machine”. Findings by Deloitte such as a fourfold rise in bar staff since the 1950s or a surge in the number of hairdressers this century suggest to the authors that technology has increased spending power, therefore creating new demand and new jobs.
So there’s no need to worry about technological unemployment, because there will always be a need for more bar-backs and haircuts? Is that bar-back better off no longer having a manufacturing job paying $40,000 per year and instead having a job paying $20,000 per year in the service industry? Is that an important job to the human species, bringing empty glasses from Point A to Point B? Is the job entirely voluntary or done out of need for income? And is this a job that just can’t possibly be done by a machine, or outright eliminated? Ever? Is the service industry really safe? Because there are already machines that can cook.
We should also probably keep in mind that the percentage of the population in the labor force peaked back in the year 2000, and has been falling since. The above chart of employment skills polarization only examines the makeup of those employed, and ignores all those unemployed and all those not even looking for work anymore. Meanwhile, GDP is still growing, so more and more work is still obviously getting done somehow despite fewer and fewer people selling their labor in the labor market.
Many Questions Unasked
Centuries of data may show us how jobs have been both destroyed and created but recent decades worth of more nuanced data more importantly show us there’s more to this story. Whenever we see someone claiming new jobs are being created and will continue to be created so as to provide everyone a job, we need to look deeper and ask the following series of questions:
What kind of job?
What are the skills required?
How much does it pay for how many hours?
Does it provide more security or less?
What are the benefits it offers?
Is the job really necessary?
Does the job provide meaning to those tasked with it?
Are jobs and work the same thing?
Is there work to do that’s more important than what the job involves?
Is working in the job actually better than not working at all?
Regarding that last question in particular, there’s this important finding which should not go ignored in any discussion celebrating job creation:
Those who moved into optimal jobs showed significant improvement in mental health compared to those who remained unemployed. Those respondents who moved into poor-quality jobs showed a significant worsening in their mental health compared to those who remained unemployed.
That’s right, having no job at all can be better than having a bullshit job. And if low-skill jobs are more likely to be worse on mental health than medium and high-skill jobs, then for decades we’ve been increasingly working in newly created jobs that are depressingly worse for us than not working in any job.
All of the above questions are important to actually ask because when we get right down to it, the mere existence of a job means very little. We have to ask additional questions about the nature of the job itself. Those not asking these additional questions are simplifying the story in such a way it becomes even simpler than a story. It becomes a fairy tale.
Yes, our economy is so far creating more and more jobs as our technology destroys old ones, but these jobs are not at all the jobs we may assume they are. They are mostly low-skill jobs, and that means mostly low-paying jobs. For every new job as a Facebook engineer, there are countless more jobs in fast food, and there are a whole lot fewer jobs in car assembly plants. There are also not as many total jobs available for everyone as we may think, or as many employers. What does this all mean to us and to our country as a whole? What does it mean for the great increases in productivity we could otherwise see if technological unemployment were instead actually our goal?
This also means that more education isn’t the answer. As the decades have passed, the population has gotten more and more educated. Our workforce now is the most educated workforce in US history, and a great deal of this education is being put to work in jobs that don’t need it, because the jobs that do need it aren’t being created in sufficient numbers. Are bar-backs with PhDs a triumph of new job creation?
The true story of technology and jobs is a story of an eroding middle, a relatively slowing top, and a vastly growing bottom.
So unless we all wish to pursue insecure lives of low-skill underpaid mostly meaningless employment thanks to all the machines increasingly doing all the rest of the work (not really for us but mostly for the benefit of those who own them), we will need to break the connection between work and income by providing everyone an income floor sufficient to both meet basic needs and purchase the goods and services the machines are providing. It’s as simple as that. Without that decoupling, there will be no economy, because there will be insufficient consumer buying power to drive it.
If we look at the details of the last few decades of job creation and destruction, we’re either going to make enough new low-skill jobs in numbers sufficient to keep unemployment numbers low enough to actually run a society... or we’re not. Either way, consumer buying power is likely to steeply erode, even after we account for the effects technology has on lowering prices because the costs of basic needs like food and housing are the costs technology has had relatively little effect on this century. Meanwhile, if we can eliminate half of our jobs in just 20 years, do we really even want to create that many tens of millions of new ways to work for someone else? Why?
There appears to be no happy ending to this story that doesn’t involve unconditional basic income. So instead of continuing to ask if jobs are going to be automated in sufficient quantities to need basic income, let’s instead start to increasingly ask if there’s any job we can’t automate so we’re all more freed to live by it.
Cool article. Hope things are going well Scott!! Thinking about organizing a conference of political radicals in Columbus, Ohio in March to focus on getting 3rd parties into the 2020 presidential debates. Would love to have you speak on BIG/UBI if you'd be free, and if the event comes together.
I've got some appearances already scheduled in March so it depends on the exact dates, but I'd be happy to speak there if all the details work out. Cheers and thanks for organizing am event like that. Good luck!
Was kinda hoping your link would go to this...
Seriously though it makes sense, even though I never thought about it that way, that medium skill jobs are the ones eroding first.
As a programmer it's always super easy to program for consistency. The more variation there is, the harder it is to code for. Something like bookkeeping is perfect for that. A human needs to learn the skill but after that it's just rote repetition... which makes it easy to automate.
Something like driving is much harder though, because of differing weather conditions and interactions with other people/drivers/animals etc. It's an easy skill for a human to learn, but hard for a computer.
BUT I think one thing missing from this analysis is that technology is now coming for those "low skill," or perhaps human-friendly, jobs. Tesla already has autopilot and the semis are coming out soon, eventually eliminating a huge swath of those human-friendly jobs. We've even already started with self-checkout lanes.
I think the full story is that while mid-skill jobs have already been eroding the low skill ones are going to start following the same pattern.
EDIT: One question: Does this in any way account for jobs lost to offshoring? I suspect a fair amount of "middle skill" jobs are now being done in India compared to being automated...
A lot of things I wasn't know is in this article XD .. good to know such things. . And you look did a very huge effort to make this.. you re just awesome man be proud !
Love the article. Great stuff here.
I was thinking that it actually makes sense for middle and high skilled jobs to get automated first. If I own a business and start looking for ways to save money though automation, I would rather eliminate the $30 per hour job first as compared to automating the $7 per hour job. Plus, automating clerical and accounting work, input etc., seems likely to be easier than automating manual labor. I used to work in a furniture store. I would open boxes and assemble furniture, put displays together, pull furniture and wrap it, and load it into trucks. I wasn't paid very much. There was a lady who took inventory and input data into the computers. She earned more though. It would be easier to automate significant parts of her job simply by using barcodes and QRs with scanners. Currently, I do a lot of input, which could easily be automated. Make a machine to automate furniture assembly and floor display setup...now try making a machine that can enter info and crunch numbers. A lot of "low skill" jobs are just harder to automate. Make a machine that flirts with customers while refilling their drinks...now make a machine for ordering and paying.
Cool to see the David Graeber link. Have you ever read "Debt: The First 5,000 Years"? Graeber was a big influence on me, pushing me in the direction of LVT+UBI and away from anarchism (ironic, since Graeber is an anarchist).
This post has received a 3.44 % upvote from @aksdwi thanks to: @scottsantens.
Sneaky Ninja Attack! You have been defended with a 1.00% vote... I was summoned by @scottsantens! I have done their bidding and now I will vanish...Whoosh
This post has received a 0.98 % upvote from @boomerang thanks to: @scottsantens
This post has received a 2.84 % upvote from @booster thanks to: @scottsantens.
I upvoted your article even though I don't agree with you on any level on the need for a basic income, because I do respect the effort and consideration you put into writing it. However I have a contrary view to yours.
In my view when you pay everyone a basic income you have to also ask the following questions.
Where does that money come from? Answer - Taxation.
By giving everyone a basic income are you increasing or decreasing the role Governments have in the lives of every person. Answer - Increasing.
By giving everyone a basic income are you going to increase the costs of goods and services just by factor of merchants charging more. Answer - Yes.
Who is going to foot the bill for the basic income. Answer - Everyone.
I know you have a lot of statistics quoted, however my experience in life has been that the only skills and opportunities that were worth having were the ones I worked for and earned myself.
No one can give you initiative. Some people will always have the initiative and drive to make their lives better despite any circumstance they've been given, and other people won't and will see it as entitlement that everyone else pays for them.
The philosophy you are advocating is one that has the cost of increasing the role and reach of Governments into the ability to dictate to every single one of us the "terms" being given alongside the basic income. How can this possibly be paid for in any way other than by Governments enforcing it? And when Governments enforce policies, such as Taxation, for example the net result is a chilling effect on individuals wanting to start and set up their own businesses.
This is just my point of view. I understand you have written this article because you clearly believe passionately in what you are writing. All I would ask you to consider are a couple of things. Is what you are proposing truly about decentralisation and human empowerment, or is it actually about Governments exerting more control? And you have missed something completely vital in this topic which is the impact that Taxation has on growth.
Maybe the next article you write could be about the effect that zero Taxation or near to zero Taxation would have on any economy. You can't really propose this "Utopian" ideal of a Basic Human Income without accepting that Taxation will have to pay for it, and therefore the side of this article that you are missing is an exploration of what would happen instead if you reduced Taxation and let People just create their own world, instead of some nightmarish centrally controlled Utopian vision.
Until you address who would actually pay for the "Basic Minimum Wage" and the fact that it would be funded through taxation, then this article is incomplete. Maybe you could make that your next article.
Please head on over to my FAQ next. This post right here is only one article talking about one thing. There you will find many more articles covering all the points you've raised in response and more.
http://www.scottsantens.com/basic-income-faq
By the way, UBI shrinks the state. It doesn't grow it. It could potentially eliminate thousands of government jobs that UBI makes entirely unnecessary with the elimination of over 100 government programs. UBI would also replace the existing controlling nature of those programs with unconditionality. Right now everything has strings. UBI cuts those strings.
Also, UBI need not be funded by income taxes, and think of UBI as a refundable tax credit. It would reduce the tax burdens of 8 of 10 households in the US.
You're so nice for commenting on this post. For that, I gave you a vote!
"Where does that money come from? Answer - Taxation."
Money for basic income come froms consent, if we so want it. Same for rights to Land. They must come from consent, or they do not exist on a moral plane at least.
"By giving everyone a basic income are you increasing or decreasing the role Governments have in the lives of every person. Answer - Increasing."
Basic income can decrease the role of government, by reducing the need to enforce property rights in Land through violence.
"By giving everyone a basic income are you going to increase the costs of goods and services just by factor of merchants charging more. Answer - Yes."
Depends. Basic income might increase the price it takes to move a man to work for another man, but only in so far as to establish that a mutual benefit is realized. A basic income can reduce prices, where people chose to work for each other more flexibly. It reduces urgency to work for landowners if adequately financed (by tapping into existing land rent), so it stands to reason that it might increase labor available for the fulfillment of the wants and needs of the bottom 80%.
"Who is going to foot the bill for the basic income. Answer - Everyone."
Everyone who enjoys titles to Land. The less Land one wishes to hold, the more likely would one be to net-benefit from the thing. As such, basic income is naturally encouraging of Commons based organization. I appreciate that to reduce dependency on centralized authority, be it in the state or in market errected snowball schemes (as much as these do transition into debt based currencies eventually anyway, till they eventually fully decouple from the base asset. That's the 'hard' currencies, at least.).
At the end of the day, we must ask: "What legitimates owning patents and physical land, what legitimates the exploitation of the biosphere?", to which we can either answer "well coming first of course" or "consent between all potentially affected actors". Or maybe other things!
John Locke might reason that leaving behind as much and as good opportunity behind for others could be good enough. ( https://en.wikipedia.org/wiki/Lockean_proviso ) Not sure how that'd be achieved, though.
Adam Smith might reason that as long as the strong state polices the market for cheapness of provision, that'd be good enough. As much as we live in times of unprecedented opportunity to take home profit margins as long as you get into a market leading role today, thanks to technology, while the competition gets nothing ( http://www.nber.org/papers/w23687 ). Economies of scale and network effect are kinda cool and all.
edit: See 16:56 for some industry survey data on supply/demand curves as they exist in reality: www.youtube.com/watch?v=MoVomqpPqto&t=16m56s
I disagree on your 3rd point, things only cost more if others aren't willing to still sell the good or do the activity at the previous price. Inflation should only be a big issue it seems if the money is printed as long as the money is not printed inflation seems like it can be ignored for the most part?
Taxation is definitely theft, but I think replacing the entire federal government with a UBI would be a grand bargain. If we could give 330 million residents a 10k per year UBI, I think most if not everyone but stock in keeping people poor (perhaps prison stocks), would drop. The US budget in 2014 I think was $3.2 trillion with an additional $600 billion for deficit spending. So UBI would be cheaper and have less government.
I'd rather have another $10k in my pocket and no war on drugs and no failed war on terror than the current Leviathan controlled by the DC cartel. Milton Friedman was an advocate of a form of a guaranteed minimum income via negative income tax that Friedman thought would help get the armies of bureaucrats out of our lives and out of our communities.
"Maybe the next article you write could be about the effect that zero Taxation or near to zero Taxation would have on any economy."
I'd be quite interested on Scott's take on this as well.
As much as abolishing taxes consequently would involve abolishing land property because that is not much more than a state enforced tax on our ability to coordinate Land use through consent, for those to benefit who have the greatest cult surrounding themselves. Heck, even without state enforced property rights, the best at building a cult around their person might as well leverage that talent to collect rent from those people who're more interested in building culture, by just defending violence with violence while talking of non-aggression (e.g. defending injust claims to the Land with violence), if they get the opportunity to do so. If they get enough cultists into their ranks, be they driven by fear of not belonging to the most militarized 'non-aggressive' faction.
Congratulations! This post has been upvoted from the communal account, @minnowsupport, by scottsantens from the Minnow Support Project. It's a witness project run by aggroed, ausbitbank, teamsteem, theprophet0, someguy123, neoxian, followbtcnews, and netuoso. The goal is to help Steemit grow by supporting Minnows. Please find us at the Peace, Abundance, and Liberty Network (PALnet) Discord Channel. It's a completely public and open space to all members of the Steemit community who voluntarily choose to be there.
If you would like to delegate to the Minnow Support Project you can do so by clicking on the following links: 50SP, 100SP, 250SP, 500SP, 1000SP, 5000SP.
Be sure to leave at least 50SP undelegated on your account.