Economics, Pethokoukis

Robots rising: The jobs challenge from smart machines

Image Credit: shutterstock

Image Credit: shutterstock

Here are MIT’s Andrew McAfee, coauthor with Erk Brynjolfsson of The Second Machine Age, and McKinsey’s James Manyika discussing automation and jobs:

James Manyika: Between the period of 2000 and 2008—2008 because that’s when the recession started, so we have a clean look—the US lost something like 5.8 million jobs in manufacturing. If you look at those jobs that we lost, only at most 20 percent of them were due to what you might call globalization, so offshoring and outsourcing. Whereas the rest of them, which is the majority, 80 percent of them, can be explained by looking at the effects of technology and the other key culprit, which is what happens to demand.

And we know that one of the things that happened in that period between 2000 and 2008 is that the demand growth for the outputs of manufacturing coming out of the US actually fell. So that was one of the big drivers for what then happened to employment. Where you had productivity growth without the demand growth, employment tends to suffer.

Andrew McAfee: There are a couple policy implications that come out pretty quickly. One is that over the longer term, we can’t rely exclusively on economic growth alone to solve all of our employment problems. Now, in the short term, economic growth is absolutely the best way to get the hiring engine kicked in again. The robots, the androids, the artificial intelligence can’t do everyone’s job yet by a long shot. So the right way in the short term to grow employment is to grow the economy. But over the longer term, it honestly feels to me like we might be in a situation where enterprises can grow and thrive and not need nearly as much labor as they’ve needed historically.

James Manyika: Certainly education will help. We know that there’s a big gap between what most economies need and what the educational training systems create to meet those needs. But I would argue even that’s not enough. So, while I think it’s comfortable for policy makers—and, in fact, correct—to say, “Let’s focus on innovation and let’s focus on entrepreneurship and let’s solve education,” those are correct, but they may not be complete answers to how we tackle employment.

So think about what Uber and its like and its kin are doing. It’s making it possible for people who have cars to suddenly turn that into a potential income-generating opportunity. Think about what models and businesses like Airbnb are doing where people can then use assets that they have, like their houses or their flats, as ways to generate income. Those are just examples of ways where, if you think about it as an income-generation question as opposed to a full-time employment problem, you expand the possibilities.

The whole exchange is worth reading. McAfee and Manyika identify two related but separate long-term problems: generating both sufficient full-time employment and income generation. Work is a good thing. It may be necessary for a good society but not sufficient. In The Second Machine Age, McAfee and Brynjolfsson quote Voltaire: “Work saves man from three great evils: boredom, vice, and need.”

Certainly many new and innovative firms, such as Facebook, generate lots of profit without many workers. And for those with jobs, what if wages by themselves don’t provide an adequate, much less rising standard of living for many people? As a New York Times story pointed out the other day, while the economy expands and jobs increase, wages are going nowhere: “…. after adjusting for inflation, real wages have increased by something like 0.5 percent” over the past year.

Wage subsidies are one option. Another is to enable workers to generate income from assets, from capital. Uber and Airbnb are one way of doing that. So is turning more workers into entrepreneurs and business owners. (Pay attention, Thomas Piketty.) So is widening capital ownership. McAfee and Brynjolfsson offer a few other ideas in their book including a sovereign wealth or national mutual fund providing a stream of dividend income to all citizens. But a key takeaway is that policies meant to boost GDP growth may not be enough to guarantee broadly shared prosperity.

Follow James Pethokoukis on Twitter at @JimPethokoukis, and AEIdeas at @AEIdeas.

5 thoughts on “Robots rising: The jobs challenge from smart machines

  1. Good piece, James.

    The story (slogan?) for the next 5-10 years will be “new ways to connect to the economy.” If we (the innovators) don’t create enough new ways for the have nots to connect to the economy, the number of under-income and non-income (forget “employment.” that’s broken forever) folks will continue to grow and eventually cripple the economy that left them behind – almost as if by some sort of (e)karmic (e)revenge.


  2. “Another is to enable workers to generate income from assets, from capital. Uber and Airbnb are one way of doing that.”

    The people who these machines are replacing do not have assets or _desirable assets_ –desirable assets should be an important caveat, especially when you cannot enumerate anything beyond living in my home (airbnb) and getting in my car (uber).

  3. As mentioned, basic income and guaranteed minimum income are possible pieces in the basket of solutions. For those wary of government control over the details, it is also heartening to consider mirror-image market variants that require the private sector to employ humans but leave the job descriptions to private innovation. More at

  4. With all due respect to those who have made careers as economists, the solution is simple. Innovation will not create jobs that machines can do, but entirely new job categories. There are only four kinds of products and services – food, clothing, shelter and entertainment. Guess which one will grow fastest in the 2nd machine age?

  5. That is distrubing. But at least the smart machines haven’t taken over spell-checking AEI’s subject lines.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>