CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press The Machines Displacing Middle Wage Jobs: Don't Let the Facts Get in the Way of a Comforting Story

The Machines Displacing Middle Wage Jobs: Don't Let the Facts Get in the Way of a Comforting Story

Sunday, 25 August 2013 10:35

We are hearing endless accounts of how technology is displacing middle wage jobs (e.g. see the piece by David Autor and David Dorn in the NYT today). That would be work like manufacturing jobs, bookkeeping jobs, and other jobs that used to provide a middle class standard of living. It's a comforting story for the people who control the media, but it happens not to be true.

The story told by Autor and Dorn is that technology displaces these jobs putting downward pressure on the wages of formerly middle class workers. At the same time it creates more jobs for the people who program the machines, hence we see higher wages for high end workers.

This story is comforting to the affluent because it means that the upward redistribution of income that we have been seeing is simply an inevitable outcome of technological progress. It might be unfortunate, but what are we supposed to do, smash the machines?

This story should strike people as absurd on its face if they are interested in anything other than a rationale for inequality. After all, how many of the winners in today's economy are actually programming the robots, as the story implies. The group of big winners includes many doctors, lawyers, and dentists, most of whom have no more computer skills than your average high school senior.

They keep their position not by mastering the technology, but rather through the old-fashion way, restricting supply. They use professional barriers and trade restrictions to limit competition. That's much easier than mastering the latest in computer technology.

This sort of abuse of market power applies to a large share, if not the majority, of the winners in today's economy. In fact, if anyone really gave a damn, they could see that the Autor-Dorn story simply does not fit the pattern of job creation that we have seen in the last decade. Their occupation analysis would show that low earning occupations have been the big job gainers since 2000. The employment share of the highest earning occupations has actually fallen slightly over this period.

However that story provides less comfort to the rich and powerful. It implies that upward redistribution is something that they did, rather than something that just happened. Therefore we will not likely see these data featured prominently in news stories and opinion pieces.


Comments (17)Add Comment
International Labour Organization Report Puts Technology in Perspective
written by Robert Salzberg, August 25, 2013 11:40
Technology is a minor player in the destruction of middle class wages. Financialization is the real major cause.

From page 52 of ILO report:

"Figure 38(a) shows that in the case of developed economies all factors contributed to the fall in the labour income share over time, with global financialization playing the largest role. The estimates mean that, in terms of relative contribution, global financialization contributes 46 per cent of the fall in labour income shares, compared to contributions of 19 per cent by globalization, 10 per cent by technology and 25 per cent by changes in two broad institutional variables: government consumption and union density. These results open up the possibility that the impact of finance may have been underestimated in many of the previous studies and suggest that overlooking the role of financial markets may have serious implications for our understanding of the causes of labour share trends."

written by skeptonomist, August 25, 2013 11:43
The kind of jobs that appeared to be making things good for the "middle class" in the US in the 50's and 60's were factory jobs and other union jobs in basic production. These jobs still exist, but now they're in other countries. Where they do exist in the US, workers have little power because of the decline of unions, so wages have not kept up with productivity.

Those who have benefited from the changes, which involve other things besides globalization, are mainly capitalists and upper managers (those who have big stock option packages), not so much professionals (most of whose services can't be outsourced). Of course finance has also swollen tremendously.

Physicians and lawyers don't control trade policy as it affects factory jobs, or the Fed, whose mandate to control inflation has involved suppression of wages. But the control of the medical profession by essentially a guild, the AMA, has probably severely inhibited the extension of modern information technology to medicine. A lot of what physicians do - mastering a huge amount of medical knowledge - should be done by computers. Training people to do this by rote is no longer efficient.
Don't really understand why you're objecting
written by Steve Roth, August 25, 2013 12:11
Seems like the article characterizes the situation quite well. While some *could* use it to justify higher earnings for the skilled/educated (cf. Mankiw's "just deserts"), that's not a real takeaway from the article as I read it.

What it fails to address is the need for policies that:

1. Ameliorate the distributional effects we're seeing, and

2. Allow for increased productivity *and* economic growth, inevitably through shifting income shares to those who spend -- removing the demand constraint on growth.

But that wasn't really the purpose of the article.

There are far better people to be arguing with...
Okey, we will try this argument again 1/2
written by Lrellok, August 25, 2013 1:06
Let me try to be more clear this time. First, I concur that technology is not, nor has ever, placed significant downward pressure on wages. Looking at equipment and technology invest as a share of GDP then comparing it to wages as a share of GDP reveals this rapidly. When Equipment and tech increases or decreases, it is followed shortly by an increase or decrease in wages.

Second, I agree that the decline in not only lower incomes but median income is the result of a constraint on supply. However, I disagree with you as to what that constraint is.

Let me construct it this way. Let us say that all farmers in the united states, when they where purchasing seeds, where obligated by contract to sell back to the seed companies, at the end of the year, 1 ton of corn seed, on penalty of forfeiture of their farm land. What would you expect to happen to the price of corn seeds? Alternatively, let us suppose that all Natural Gas wells had to be operated at maximum capacity, on penalty of forfeiture of drilling rights? What would you expect to happen to natural gas prices?

There is a phenomenon that for arguments sake i am calling a "quantity bound", that for some reason is totally unrecognized in the field of economics. I have found not one paper written on quantity bounds anywhere on lexus nexus or any google search. Yet inherent to the idea of free markets is that if prices are to correct efficiently, then individuals must be able to start or stop buying or selling as prices change, in order for prices to adjust correctly.
In our society, all individuals of working age and able body, particularly all males, are obligated on pain of starvation, to sell 40 hours per week of labor, or as close to 40 hours as the will be allowed. This effectively means that the quantity of labor supplied to the market is always has a binding floor of the number of working age males times 40 hours. Now that women of working age are expected to work 40 hours as well, this binding floor has increased to all males plus most females.

Please, LOOK AT THE GRAPHS SIR. Between 1968 and 1988 Labor force to population (16+) rises from 60% to 66%, while wages as a share of GDP fall from 52% to 46%. A 6% increase in labor force produced a 6% decline in wages/GDP. That does not seem complex to me, that does not seem complex at all. If we extend the data even further back, using the 1970 census report, we can see that at no point since the 1900s has Labor force ever totaled more then 42% of the total population (0+) until this sudden climb beginning in the 70's.
Further, if that where the result of a quantity bound, then it would be entirely predictable that wages would continue to fall, as quantity supplied is unable to correct to the new lower rate, prices would continue falling waiting for quantity for correct.
Let me try this again 2/2
written by Lrellok, August 25, 2013 1:13
Every examination of historical data i have seen shows that the separation of wages from production increases began around 1970, exactly when the LF increase started. I ask, please, explain to me why supply and demand should not function in wages, and if they do not, why are we using them to set wage prices?
It is not, in my view, that doctors, or lawyers, or public sector unions are over-payed. That is a complete inversion of the case. It is that everyone else is underpaid, forced into competition in a marketplace that is systemically over supplied due to a societal bound placed on quantity.
This explains nearly everything that has transpired in our economy since. First, Stock prices are set based on the profits returned, but the profits where distorted up due to artificially low wages, thus the inflated prices where unsustainable and severe crashes occurred. Second, because wages where to low, business ventures no one in their right mind should have looked at suddenly became profitable, and wide scale miss-allocation of capital ensued. This explains all the dot come companies that thought they could make money with no products.

Lastly, no i am not being sexist. I will happily advocate for 12% of all men to sit in their apartments playing video games if women are willing to pay for them to do so, and if paid to do so would happily join them. It is glaringly apparent that the private sector of the united states cannot support more then 50% LF to population (16+) while maintaining market efficient wages. We could of course expand public sector jobs to make up the difference, like we did between 1947 and 1972, from 5.47% to 9.24% of population (16+). If anything, the fact that the steady increase in public sector employment stabilized wages indicates that this dynamic of private sector miss-pricing labor by oversupply is not only systemic its historical. Which gets us rapidly into the territory of "Most of Marx's questions where right, most of Marx's answers where wrong". But that can be argued another day.
Saw this earlier and thought it was silly
written by Alex Bollinger, August 25, 2013 2:16
People have been saying this for at least two centuries: new technology is destroying all the good jobs! Marx thought unemployment would grow monotonically as machines replaced laborers... somehow people thought up new stuff they could consume to make up for productivity increases.

For example, farmers were 90% of the labor force in 1790 in the US. Now they're around 3%. So all those farming jobs have been destroyed! That explains our 87% unemployment rate! Um, no.

If anything, machines should (the moral should, not the positive should) make workers' lives better. Your job got replaced by a machine? Awesome. Now you don't have to work. Work sucks. We just need to keep the paycheck going by, say, taxing the employer.
Between the IMF, newsmedia, and Jackson Hole...
written by JaaaaayCeeeee, August 25, 2013 2:35

Dean Baker has been an oasis of sanity today. The propaganda is really getting rammed down a lot of throats hungry for facts, if not food.
written by Tom, August 25, 2013 4:39
I didn't read the article that Dean is writing about and I agree that overall technological changes don't have much impact on employment. Now having said that, what I see to contradict that to some extent is every time I walk into my local Safeway. They use to always have 6 registers open. That of course requires 6 cashiers. Now I go in and they have 2 registers open and manned and 4 self check outs open manned by one person. Seems to me Safeway by using these self check out scanners has reduced by 50% the number of cashiers they use to have. That's at least 3 jobs and probably more if you go second shift that no longer exist at that Safeway.

Now, since Safeway is buying those machines maybe the manufacturer is hiring more people to make them and their is no overall net loss to employment, but try telling that to the cashiers that are not being hired or having their hours reduced or being layed off.
Dean is right for now, but....
written by David Cay Johnston, August 25, 2013 5:06
... in the long run is what is true today -- that fictionalization, lack of regulations and the destruction of unions are the real issues.

But physical capital is in fact doing ever more labor.

We are in the early stages of machines doing work. Look at the advances in manufacturing robots since the GM Lordstown plant opened just 40-some years ago.

As we become more adept at using digital and even biological machines we will many current jobs go away, including service jobs.

While I agree with Dean's major point FOR NOW, in the lifetimes of many people breathing today we will see many existing jobs replaced because of our growing mastery of the physical and structured universes. And if ownership of physical capital is highly concentrated so, too, will incomes be more concentrated at the top if we do not adapt.

It's Not the Robots Stupid: Chicago School is Right for the Wrong Reason
written by Last Mover, August 25, 2013 5:39

@Alex Bolinger
For example, farmers were 90% of the labor force in 1790 in the US. Now they're around 3%. So all those farming jobs have been destroyed! That explains our 87% unemployment rate! Um, no.

Exactly. And think Malthus as well in the opposite direction, that mankind was doomed because population growth would outgrow the productivity necessary to live above subsistence levels. Wrong again.

At a minimum the robots-are-taking-jobs scare would depend on massive market failure to happen, failure of a nature that directly contradicts the assertion itself.

Specifically, from a condition of full employment, demand must fall off dramatically somewhere for jobs to be "taken by robots", otherwise jobs lost to robots would be absorbed into both, robot related jobs as well as other jobs ultimately created by the added productivity value of robots when translated into demand.

Further, with both full employment and competition (the lack thereof described by Dean Baker above), it doesn't necessarily follow that new jobs related to (new) robots are superior in wages and benefits to the jobs they replace (and ultimately leads to even more inequality). At full employment replaced workers could easily end up better off in different jobs, robot related or not, as did many replaced farmers in an earlier era.

In isolation, the claim that robots will take jobs makes no more sense than saying desktop computers took jobs starting in the '80s, which did not cause aggregate demand to fall off, nor were computers themselves responsible for the increase in concentrated income and wealth (it was market structure instead). Computers did cause sharp cost and price reductions in selected areas resulting in heavy substitution effects, but "taking" net jobs to get there was not necessary under full employment.

Two kinds of market failure are taking jobs today unrelated to robots (or interpreted differently, related to robots in a way that makes them complicit in the market failures).

The first failure is the overwhelming anti-competitive power to dictate market prices and quantities rather than have them actually set by "free markets" (or if set by competitive markets, then the same powers capture the gains rather than consumers anyway, as did the MNCs with manufacturing).

The second failure is the inability of the private sector to restore the economy to full employment at the macro level (despite having caused the great recession as well).

Both failures were and remain instrumental in redistributing income and wealth gains to the rich and ultra rich away from the middle class. Neither failure depends on robots in any way.

In fact, given claims by the Chicago School on how prices should flex downwards to restore aggregate demand, robots should accomplish exactly this result as they increase productivity, creating enough demand to keep everyone employed in jobs that pay even more after being replaced by a robot.

Of course the Chicage School conveniently omits the more obvious cases readily documented, say in health care for example, that advances from robots are not allowed to bring prices down because of market power, so the two market failures that are actually taking jobs never go away anyway.
Last Mover, a bit more on that please.
written by David Cay Johnston, August 25, 2013 6:36
You wrote:

"In fact, given claims by the Chicago School on how prices should flex downwards to restore aggregate demand, robots should accomplish exactly this result as they increase productivity, creating enough demand to keep everyone employed in jobs that pay even more after being replaced by a robot."

You had me up to "employed in jobs that pay even more...."

Freeing people from farming enabled economic growth. And the legal rules we apply shape economic results -- market power, asymmetrical knowledge and rights, etc.

But what happens when we need, say, only 10 percent of the current global manufacturing workforce to make things?

You seem to assume productivity gains will somehow benefit actual workers as opposed to the owners of those devices whose name derives from a Hungarian word that means slave.

What happens if the productivity increases from robots flow only, or overwhelmingly, to the owners of these machines?

The ownership of both physical and financial capital already being highly concentrated and if you look at controlling ownership being even more highly concentrated. In a world where mechanical slaves are much more efficient and productive than today -- and may even build, maintain and repair one another -- what happens to the incomes of people with labor to sell, but little or no capital?

written by Kat, August 26, 2013 5:40
Now, since Safeway is buying those machines maybe the manufacturer is hiring more people to make them and their is no overall net loss to employment, but try telling that to the cashiers that are not being hired or having their hours reduced or being layed off.

But I don't think that is true. There can be no way that the net effect on employment is neutral. I guess another question to ask is is the cost savings passed on to customers?
I read an article about the increasing costs of college and one reason given was the use of "state of the art" classrooms. I'm assuming they meant larger classrooms with technology replacing some actual human instructors. Why should this drive up the cost? Why do we accept such explanations?
written by Last Mover, August 26, 2013 6:09

@David Cay Johnston

Good points and agree other than predicted results of technological advance must benefit owners over others. Your excellent work has demonstrated repeatedly on an empirical level how the ravages of private market power have plunged America backwards in many ways, even as its GNP grows albeit way below potential.

Since it is about the legal rules applied that shape the economic results, that's why the pillaging and plundering by those who own the means of production can't be blamed on robots or any other technology per se.

Those rules allowed displaced farmers a better standard of living then, that they would never see today where unlike farmers, the middle class has stagnated in its tracks and is actually worse off from the dramatic changes in the rules starting in the '70s.

So the essential economic question is, can the legal rules be changed to break up the monopoly power of ownership and control while adopting the most productive technology available to maximize growth at the same time?

The essential corollary question is, once "free markets" are actually legally shaped to be competitive when they can be with enough players - or regulated enough to benefit consumers - not owners - from scale and scope economies when they can't be competitive, will evolving technology like robots act to maximize the size of the economic pie and distribute benefits in ways that match earned gains with value added at the same time?

In short, is there something inherent in the nature of increasing efficiency from technology itself that requires more concentration of income and wealth, and increases unemployment at the same time? If so, this implies markets can't work at all regardless of the rules, and this is how critics of the anti-robot crowd frame it as socialist luddites blocking economic growth.

The answer is no, technology per se is not the problem. A good example is how Dean Baker constantly reminds everyone why pharmaceuticals from Big Pharma need not be priced hundreds of times above the marginal cost of production they would sell for under effective competition.

That low cost of production at the margin is due to technological advance. The extortion prices are due to legal rules, not the technology. Dean Baker's point is that different rules could easily be applied to rid the market of extortion prices while preserving efficiency gains from technology and make them available at prices still sufficient to cover cost of production.

Or consider more generally the health care systems of other developed countries at half the cost of America, where they learned long ago how to employ technology efficiently to benefit the most rather than the few, without sacrificing incentives necessary for technology to evolve. (Note you have made similar comparisons on broadband for example.)
Hey, everybody, David Cay Johnston is here!
written by David M, August 26, 2013 9:29
Big fan.

But I do want to respond to a couple of items in your first post. First, "labor" and "jobs" shouldn't be conflated. If the standard work week dropped from 40 hrs to 20 hrs per week (preferably by worker action and changing social convention rather than government mandate), there could be more jobs even if less labor needed to be performed. Yes, I know the lump of labor fallacy arguments, but I don't think you can deny that if we were still working the 60-80 hour weeks that were common 100 years ago, there would be more unemployment.

Also, don't overlook the ability of people to create new fields of consumption. I'm not talking about the over-hyped entrepreneurs, but the immense expansion of intellectual and creative realms. Many people may not like academic disciplines like Women's Studies or genres like gangsta rap, but in fact both of those opened up new areas of consumption and created thousands of jobs with very little cannibalization of existing disciplines or genres, and very little resource usage or environmental impact.
written by NWsteve, August 27, 2013 12:11
one of the *better* comments-section in this location in a long time...

thanks everyone--very stimulating and informative...
Autor and Dorn's unexamined premise
written by Sandwichman, August 27, 2013 1:10
I sent an email to David Autor and David Dorn criticizing their NYT piece and received a gracious and thoughtful reply from Autor. I posted the text of my original email at EconoSpeak http://econospeak.blogspot.com...ected.html, so I won't repeat it here.

Here is my response to the reply from Autor:

Dear David Autor,

Thanks so much for your thoughtful and prompt reply. I appreciate your purpose to, as you say, "give a more nuanced interpretation of the legitimate concerns surrounding the impact of rapid technological [change] on job opportunities." My point was actually that invoking the lump-of-labor canard poses a hindrance rather than a help to achieving that laudable objective, which I share. As I wrote, my intention was not so much to dispute your arguments about technology and employment or even about the lump-of-labor notion itself as to bring to your attention the positive contribution that could be made by giving a fair hearing to the actual views of those who are worried about technological unemployment.

I didn't "miss" your point at all, instead I chose to avoid piling a gratuitous critique of your conclusion on top of my main criticism of your premise. But since you asked... I would characterize your outlook and prediction as fitting neatly into Keynes's category of "too easy, too useless a task if in tempestuous seasons they [economists] can only tell us that when the storm is long past the ocean is flat again." Your "too easy, too useless" outlook and prediction flows seamlessly from the unexamined premise of viewing labor as a commodity. Karl Polanyi argued that such a description of labor is "entirely fictitious" yet actual markets are based on this fiction.

An alternative description of labor is as a commons, or to use the late Elinor Ostrom's term, a "common-pool resource." The words and actions of ordinary workers, trade unions and even machine-breakers make a great deal more sense from the perspective of treating labor as a common-pool resource rather than as a commodity. By contrast, the inane assumption attributed to workers by their detractors is itself predicated on the unstated assumption of the unquestionable commodity status of labor. Profoundly different policy implications flow from the two contrasting assumptions, as do fundamentally different predictions about the future. It seems to me that a "more nuanced interpretation of the legitimate concerns..." would seek to include both the common-pool resource and the commodity interpretations of labor rather than to exclusively feature the latter while inadvertently disparaging the former by attributing it to a belief in a spurious fallacy.

I'm only scratching the surface here. I could go into much more detail and provide extensive reference on the question of viewing labor as a commodity versus viewing it as a common pool resource but life is short and I don't want to annoy you with unsolicited "singing lessons."


Tom Walker
Executive Director
written by Gary Reber, August 27, 2013 3:34
Of course, not EVERY PERSON will be replaced by technological progress evident in advanced human-intelligent machines, super-automation, robotics, digital computerized operations, etc., but companies will strive to keep labor input and other costs at a minimum. As a result, private sector job creation in numbers that match the pool of people willing and able to work is constantly being eroded by physical productive capital’s ever increasing role. The potential is the ability to build a FUTURE economy that can support general affluence for EVERY citizen by unleashing the full productive power of technological innovation and invention to free, where ever possible, humans from toil labor and provide the equal opportunity to acquire ownership in FUTURE wealth-creating, income-generating productive capital assets.

These NYT authors need to understand that human productivity has not advanced, but that the productiveness of the non-human factor of production––productive capital––is the reason that private sector corporations, majority owned by the "1 percent," are utilizing the non-human factor of production increasingly to create efficiencies and save labor costs. It is the function of technology to save labor from toil and to enable us to do things that otherwise is humanly impossible without non-human input. The critical question becomes who should own productive capital? The issue of OWNERSHIP is unbelievably overlooked by those in academia and politics, as well as authors such as Autor and Dorn. Yet we live in country founded upon private property rights.

Today, large streams of data, coupled with statistical analysis and sophisticated algorithms, are rapidly gaining importance in almost every field of science, politics, journalism, and much more. What does this mean for the future of work?

With increasing punditry, scholars and others are writing about the impact of the Second Industrial Revolution where tectonic shifts in the technologies of production are destroying and degrading jobs due to the shift from labor worker input to the non-human factor––human-intelligent machines, super-automation, robotics, digital computer operations, etc.

The question that requires an answer is now timely before us. It was first posed by binary economist Louis Kelso in the 1950s but has never been thoroughly discussed on the national stage. Nor has there been the proper education of our citizenry that addresses what economic justice is and what ownership is. Therefore, by ignoring such issues of economic justice and ownership, our leaders are ignoring the concentration of power through ownership of productive capital, with the result of denying the 99 percenters equal opportunity to become capital owners. The question, as posed by Kelso is: “how are all individuals to be adequately productive when a tiny minority (capital workers) produce a major share and the vast majority (labor workers), a minor share of total goods and service,” and thus, “how do we get from a world in which the most productive factor—physical capital—is owned by a handful of people, to a world where the same factor is owned by a majority—and ultimately 100 percent—of the consumers, while respecting all the constitutional rights of present capital owners?”

See "Financing Economic Growth With 'FUTURE SAVINGS': Solutions To Protect America From Economic Decline" at NationOfChange.org http://www.nationofchange.org/...-137450624

Write comment

(Only one link allowed per comment)

This content has been locked. You can no longer post any comments.


Support this blog, donate
Combined Federal Campaign #79613

About Beat the Press

Dean Baker is co-director of the Center for Economic and Policy Research in Washington, D.C. He is the author of several books, his latest being The End of Loser Liberalism: Making Markets Progressive. Read more about Dean.