The Weekly Sift

Not what I’m talking about

No, I don’t mean the Cylons from Battlestar Galactica. I’m talking about what the economists call technological unemployment, i.e., there are no jobs because machines do everything. Or, as Keynes defined it:

unemployment due to our discovery of means of economizing the use of labor outrunning the pace at which we can find new uses for labor.

Now, this is not a new idea, so it’s something of a mystery why so many people started talking about it this week. I think the original impetus was the news that Apple is planning to move some of its Mac production back to the U.S. This validates the predictions of a rebound in American manufacturing that have been bouncing around for a while now, and fits with the “insourcing” trend identified in the current issue of Atlantic.

After the initial cheering died down, people started envisioning those new manufacturing jobs: Probably there won’t be many of them, because even minimum-wage Americans make a lot more than Chinese factory workers. So any large-scale new American manufacturing plant only makes sense if most of the work is done by entities not covered under minimum-wage laws or represented by unions: robots, in other words. (Another workerless manufacturing technology is 3D printing. It’s going to be huge someday, but it doesn’t seem to be news on any particular day.)

So the complete story is: High-paying unionized factory jobs in the Rust Belt get replaced by low-paying non-unionized factory jobs in China, which in turn get replaced gradually by robots in China, until eventually there’s no reason not to have those robots work in the U.S., closer to the buyer.

As more robots are built, largely by other robots, “assembly can be done here as well as anywhere else,” said Rob Enderle, an analyst based in San Jose, Calif., who has been following the computer electronics industry for a quarter-century. “That will replace most of the workers, though you will need a few people to manage the robots.”

So if the “insourcing boom” isn’t going to be a jobs boom, where will the jobs of the future come from? That was the theme of Race Against the Machine and The Lights in the Tunnel, two books I already told you about last year. But this week the topic got hot again. Kevin Drum worried about it. And Matt Yglesias said, basically: No problem, when the robots do all the manufacturing, we’ll still have jobs taking care of old people, an answer that Drum, in turn, derided as “our bedpan and canasta future“. Paul Krugman also weighed in, twice, three times.

Krugman’s insight was the most interesting: Moving jobs to China (or back) just replaces labor with other labor. But robots and other automations replace labor with capital. (Or, if you want to use Marxist terminology that Krugman avoids: It replaces living labor with dead labor, since capital is just the residue of labor done in the past.) This could be why the corporate-profit share of the economy is surging while labor’s share of the economy is falling.

If this is the wave of the future, it makes nonsense of just about all the conventional wisdom on reducing inequality. Better education won’t do much to reduce inequality if the big rewards simply go to those with the most assets. Creating an “opportunity society”, or whatever it is the likes of Paul Ryan etc. are selling this week, won’t do much if the most important asset you can have in life is, well, lots of assets inherited from your parents.

Kevin Drum plays this vision out:

the owners of capital will automate more and more, putting more and more people out of work. Liberals will continue to think that perhaps this can be solved with better education. Conservatives will continue to insist that people without jobs are lazy bums who shouldn’t be coddled. The lucky owners of capital won’t care. Their numbers will decline, but the ones who remain will get richer and richer. The rest of us will have no jobs, and even with all this lovely automation, our government-supplied welfare checks will be meager enough that our lives will be miserable.

I don’t have an answer, either to jobs or inequality, but there are two general observations you need to keep in mind when you think about this stuff:

First, technological unemployment is fundamentally a social problem, not an economic problem. For comparison, think about all the automation that has replaced “jobs” inside a parallel social system, the family. Our ancestors used to spend an enormous amount of time walking down to the creek and carrying back buckets of water (which get damn heavy after a few steps). Modern plumbing delivers that water to your kitchen with the turn of a tap. Is that a problem? No. Within the family, replacing work with leisure is a pure benefit.

The problem only shows up in the money economy, where “leisure” becomes “unemployment”. In a world of robot-produced abundance, our system for distributing the abundant goods gets screwed up. The economy no longer needs human workers to produce goods, but we individual humans still need jobs to justify our claim to receive those goods. An outside observer might wonder: Couldn’t those goods just be given to the people who need or want them? Answer: Not without screwing up our motivational systems and our ideas about personal worth and identity.

It’s a social problem.

(As an aside, this is why supply-side economics looks at things exactly backwards: Increasingly, supply is not the problem. Demand is the problem. How do we give people permission to consume the goods we can so easily produce? Agriculture, where technological unemployment started, is a prime example: The world produces plenty of food and could produce a lot more. But hungry people have no money; that’s another way of saying that our economic system can’t justify distributing our abundant food to the people who need it.)

Second, whenever you visualize this process, you can’t forget the time dimension. Economists love to talk about how this all works out in the long run: When you no longer need to spend your time hauling water, you realize that you’ve always wanted to see Paris. As the old goods become producible with little labor, desires for new goods arise, creating new jobs — airline attendants, tour guides, museum curators, waiters in sidewalk cafes. My water-hauling ancestors had no use for such people, but I do.

The real problem, as Keynes already knew in 1930, is pace. It’s not that there will never again be anything of economic value for humans to do. But will new desires arise and new markets develop as fast as automation replaces current jobs? Keynes’ most famous (and most inappropriately quoted) line is “In the long run we are all dead.” He was making precisely this point about time: Don’t give me your steady-state model and claim that the real economy will “eventually” converge to it. If you’re not worrying about the timing, you’re not working on the real problem.