The future of work in the second machine age is up to us
“Big Thinkers” about the role of technology in the U.S. economy are roughly divided into two camps when it comes to the consequences of rapid technological change on the U.S. workforce. There is the techno-optimist view that better technology complements workers and hence benefits them by raising wages. And there’s the pessimistic view that better technology substitutes for workers and therefore displaces and harms them. A debate between the two views was probably what the organizers intended for an event last week hosted by The Brookings Institution’s Hamilton Project entitled “The Future of Work in the Age of the Machine.”
The impetus for the forum was the influential 2014 book “The Second Machine Age” by professors Erik Brynnjolfsson and Andrew McAfee at the Massachusetts Institute of Technology. The authors argue that increasingly “smart” technology displaces workers by reducing the range of tasks that require human ingenuity, and by enabling economic arrangements such as off-shoring that rely on instantaneous global communication and replicability. Brynnjolfsson and McAfee are clearly in the pessimists’ camp.
Until recently, economists were largely in the optimist camp. Sure, some jobs—think buggy whip manufacturers, typists, or travel agents—might disappear, but others would arise to take their place. In the long run, increased productivity would benefit everyone in the form of higher wages.
Yet the debate last week actually highlighted a third position. If either the techno-optimists or the techno-pessimists are right, then we should see a major positive impact on worker productivity. But it just isn’t there in the data. If anything, the rate of technological change in the United States has decreased since at least 2003, specifically in the technology sectors widely thought to be most innovative.
In contrast, we definitely see worker displacement, stagnant earnings, a failing job ladder, rising inequality at the top, “over-education” (workers taking jobs for which they’re historically overqualified), and declining rates of employment-to-population and household and small business formation. What we do not see are the productivity gains, either on a micro or macro level, that are supposedly driving worker displacement. (See Figure 1.)
Former Treasury Secretary Larry Summers made this point forcefully at the Hamilton Project event. He said “people see there’s already a lot of disemployment but not a lot of productivity growth.” And he continued by asserting that “the core problem is that there aren’t enough jobs,” and that it’s hard to believe the future promise of labor-supplanting technology is driving current displacement. The reason, he said, is that we’d expect to see the installment of new labor-saving systems that would cause a temporary increase in labor demand during the transition.
Summers noted that back when he was an undergraduate at MIT in the 1960s, his professors said labor would not be displaced by technology. In those days, the non-employment rate for prime-age male workers was 6 percent. Now it’s 16 percent. Summers’ co-panelist David Autor added that since 2000, the education wage premium has reached a plateau and the rate of over-education has increased, both of which are hard to square with the argument that the reason for rising inequality is the advance of technology. Summers added that the idea that more education solves the problem of displaced labor is “fundamentally an evasion.” Summers’ arguments and Autor’s observation imply that if we’re wondering how things got so bad for workers, it’s not because we live in the Second Machine Age.
So if not technology, what explains labor displacement?
Broadly speaking, the explanation is this: market practices and public policies that favor managers over workers, and those who make their living by owning capital over those who make their living by earning wages. That choice lurks behind the decline in full employment as a priority in macroeconomic policymaking. It’s also behind a shift in the legal standards, mores, and incentives of corporate management in favor of the interests of owners over other stakeholders. That choice is also evident in the abandonment of long-term productive investment as a priority in public budgeting in favor of upper-income tax breaks and retirement programs for the elderly.
As Summers noted at the Hamilton Project’s event, there seems to be a lot of so-called rents—economics speak for excessive payment for something beyond its actual value—in corporate profits that can’t be understood as the fruits of productive investment. The big question is: who gets those rents? In 1988, Summers wrote an article fleshing out the idea that the division of rents between corporate stakeholders is what drives rising inequality. More than a quarter century later, he could not have been more prescient.
The good news is that if such a profound shift played out over only three or four decades, then it’s reversible. That wouldn’t be true if it were the result of the technological trends detailed in “The Second Machine Age.” So what should be the focus of public policy is to figure out ways for workers to accrue more of corporate earnings, for more unemployed and underemployed people to find full-time, productive jobs, and for the broader economy to serve the interests of the actual people who inhabit it—those who overwhelmingly derive their living from their labor.
We know what needs to be done and how to do it, because we’ve done it before. (See Figure 2.) But it’s a lot harder to actually do than doubling the number of logic gates on a computer chip every two years—the ostensible tech explanation for our current economic woes.