CEPR - Center for Economic and Policy Research


En Español

Em Português

Other Languages

Home Publications Blogs Beat the Press Will Robots Help Your Employment Prospects? Autor and Yglesias

Will Robots Help Your Employment Prospects? Autor and Yglesias

Monday, 25 August 2014 04:08

The usually astute Matthew Yglesias falls off the track with his discussion of David Autor's latest paper on technology and wages. For background, Autor is the guru of the job polarization story: the idea that technology is destroying middle-paying jobs leaving only those at the top and bottom. He presented a new paper at the Fed's annual conference at Jackson Hole which reassesses his prior work.

Matt's take on this paper has Autor telling us that robots may not take our jobs, but they will cut our pay. That isn't the story as I see it. Technology always devalues some jobs while increasing the productivity and wages of other jobs (that's why average wages are higher today than they were 100 years ago). New technologies like robots will not be different in this respect.

What's new and newsworthy in the Autor paper is the acknowledgement that his occupation story really cannot explain trends in wage inequality. Here's a figure from Autor's paper that Matt uses in his post.


Note that there is no job polarization in the period 2000-2007 and only very modest high end job growth in the period 2007-2012. The main story in these periods is the growth in the share of low-end occupations. Yet we continued to see a sharp increase in high end wages relative to everyone else.

This is a problem not only for the post-2000 period, but for the whole period. If high-end wages increase relative to other wages when their occupation share is not rising in the period 2000-2012, why would we think that the mix of occupations explains wage trends in earlier periods? And of course the sharpest increase in shares is for occupations at the bottom end of the skills distribution, the workers who have seen the sharpest drop in relative wages in the years since 2000 as well as the longer period since 1979.

In other words, there is no link between changes in occupation shares and wage trends, a point that my friends and colleagues, Larry Mishel, John Schmitt, and Heidi Shierholz, have been making for several years. These points in Autor's new paper appeared in their paper, Don't Blame the Robots. (Autor was a discussant of an earlier version of this paper at the American Economic Association meetings in 2013, although it is not cited in his new paper.)

Anyhow, given the eagerness with which the punditry embraced Autor's hollowing out of the middle story, the fact that he has now moved away from it should be big news. This means that the economics profession does not have a way to blame the growth of wage inequality on technology. And if it wasn't technology that gave us inequality, then we might start thinking about policy. 


Comments (5)Add Comment
written by LSTB, August 25, 2014 6:18
Autor may've moderated some of his views about job polarization, but according to Yglesias he apparently still thinks the solution to the problem of Eloi lawyers and Morlock material movers is to send the Morlocks to law school and then deny the existence of any material movers who happen to have law degrees.

If the research findings change but the policy prescription stays the same, then I'm not seeing much progress. We can expect Autor to still be heralded by the college-for-all crowd that have replicated their way into the media.

(As for Yglesias, I don't think he's particularly astute.)
Desktop Computers Did Not Take Your Jobs America - Predators Against Free Markets Did
written by Last Mover, August 25, 2014 7:40

If there is no link between wage and occupation at all, then the labor market is a total failure. Wage as the price of labor would have nothing to do with providing it as a scarce resource from the supply side or allocating it on the demand side.

That the relation between changes in occupation shares and wages has broken down serves as a 30-year marker when the economic predation of America began to transform it into the oligarchy it is today.

The redistribution of productivity gains to the upper 1% and especially the .1% and .01% is directly related to the breakdown between occupational titles and associated skills.

One way to understand this is observe how many of the technical "skills" in question are easy to learn on the job. Increasingly unrelated to prior education and training, the requirements are more basic, related to things like good hand-eye coordination.

If the pay is good the selection of employees will be based heavily on other non-skill, non-market factors. If the pay is low employers will feed from the bottom of the skill bucket to still acquire the necessary skills from anyone available and desperate enough.

The other primary category of "skills" that drives wide variations in pay within occupational categories has evolved into "skills" carefullly honed to defeat markets, rather than respond to them.

For example tens of thousands of lobbyists get paid millions to successfully undermine "free" markets in every way imaginable. The result should not be a surprise - occupational categories increasingly less related to any skill at all directly related to added value productivity.

Then is a third category where markets - even when allowed to work - have long proven to fail in the first place, where "skills" are necessarily priced based on non-market factors.

Traditional examples were police and national defense resources, but even this has broken down under the pillage and plunder of predator corruption to monetize and privatize every "public good market failure" in sight ... to guarantee its failure of course.

The robots didn't take jobs anymore than desktop computers did. The predators took them, and they took them in the name of the very free markets they preach for everyone else but themselves:

Wanted, Highly Skilled Employee of Visual Imagery Who Believes Free Markets Reduced Your Pay to a Subsistence Wage With No Benefits - Literate Employees Who Can Read Need Not Apply
Same old Autor
written by Sandwichman, August 25, 2014 10:04

"If the research findings change but the policy prescription stays the same, then I'm not seeing much progress."

Yes, I've really surprised at your elevation of Autor's paper to breakthrough status, Dean. The data analysis may change but the framework and the panacea remain the same:
"Thus, human capital investment must be at the heart of any long-­??term strategy for producing skills that are complemented rather than substituted by technology."

Awful graph!
written by Melissa, August 26, 2014 6:06
This has to be one of the worst graphs I've ever seen. I don't know how common it is to represent change over time using overlapping lines, but I've never seen it before and frankly can't make sense of it. And I'm pretty numerate. I wonder what Edward Tufte would say of this?
Cardiff Garcia on Autor at Alphaville FT
written by Sandwichman, August 29, 2014 11:20


"Stated plainly, the U-­shaped growth of occupational employment came increasingly to resemble a downward ramp in the 2000s.

"Despite these recent trends and the foreboding pace of automation, Autor is relatively sanguine about the longer-term prospects for the labour market.


"Well, here’s hoping.

"Autor’s paper is excellent on the recent past, but it’s difficult to have much confidence in his prediction for the future. The pace and impact of technological progress is simply unknown.


"The appeal to history isn't pointless, but neither is it all that persuasive given that the relevant history is only 250 years old."

Garcia nails it. This is why I find Dean's take so puzzling. Dean is so dazzled by the empirical truth-telling that he doesn't notice Autor is using it as a scaffold for the same old same old nostrum.

The one quibble I have with Garcia's take is to point out that Autor's "appeal to history" is profoundly, profoundly unhistorical. It's a Fractured Fairy Tale history.

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.