Better Living Through Robots

Economists debate whether automation and technology are the way out of a productivity rut.

A double-arm prototype robot moves a box during a media preview hosted by Hitachi Transport System at a warehouse in Noda City, Japan, on Aug. 25, 2015.

Photographer: Akio Kon/Bloomberg

Robert Gordon, an economist at Northwestern University, likes to play a game he calls Find the Robot. As he goes about his everyday life—shopping, traveling through airports—he looks for machines performing tasks that humans once handled. Most of what he sees doesn’t impress him. ATMs, self-checkout kiosks, and boarding pass scanners have been around for years. Beyond that, not a lot has changed. “It’s very hard for robots to do things that are extremely ordinary for humans,” Gordon says. “Turns out that teaching machines to do something like folding laundry is almost impossible.”

IBM’s supercomputer Watson made history four years ago when it beat two contestants on the quiz show Jeopardy!, yet for most of us artificial intelligence is still the stuff of sci-fi movies. And even though there’s lots of talk about drones and driverless cars, how close are they really to changing how we live? To listen to Gordon, not very. “I see stasis everywhere,” says the scholar, whose new book, The Rise and Fall of American Growth, posits that the U.S. economy’s best days are behind it. “Stasis in the way offices work, stasis in the way retailers work, and stasis in the way factories produce goods.”


Gordon’s pessimism stands in sharp contrast to the upbeat theme of this year’s annual meeting of the World Economic Forum in Davos: Mastering the Fourth Industrial Revolution. It suggests we’re on the cusp of a transformative era that will rival those triggered by the introduction of the steam engine, electrification, and computers. “The speed of current breakthroughs has no historical precedent,” proclaimed WEF founder Klaus Schwab in a December post on the website of Foreign Affairs, in which he name-checked a slew of emerging technologies, including 3D printing, quantum computing, and autonomous vehicles. “When compared with previous industrial revolutions, the Fourth is evolving at an exponential rather than a linear pace.”

Apparently it’s moving so fast that it has yet to make a mark on economic data. The one place where you’d expect to see the impact of all of these advances is in labor productivity, which in the U.S. at least has been mired in a decade-long slump. Over the course of history, breakthroughs such as electrical motors and computers have enabled workers to produce more goods and services than they otherwise could for a given number of work hours. Annual productivity growth in the U.S. averaged 1.5 percent from the first quarter of 2008 through the same period in 2014. That’s less than half the 3.5 percent rate during the previous boom, which lasted from 1996 to 2003.

The data have become fodder for a lively debate about the potential of new technologies to lift growth. On one side are pessimists such as Gordon and Tyler Cowen, an economist at George Mason University and the author of The Great Stagnation, a 2011 best-seller whose thesis is summed up in its subtitle, which in part reads: How America Ate All the Low-Hanging Fruit of Modern History. Another member of this camp is John Fernald, an economist at the Federal Reserve Bank of San Francisco who’s considered an expert in measuring technology’s contribution to productivity. In a 2014 research paper, Fernald made the case that even though IT industries contributed greatly to the productivity boom of the late ’90s and early 2000s, IT-related productivity then slowed dramatically. “If you look at the future through the lens of the past, it’s very hard to see anything other than continued incremental, modest gains,” Fernald says. In other words, productivity booms are more the anomaly than the norm. “We’re simply reverting back to a lower mean,” he says.

On the other side are optimists best represented by Erik Brynjolfsson and Andrew McAfee, the authors of The Second Machine Age, who are confident that another golden age lies ahead. The two argue that it takes time for methods of production to adapt to new technologies, which is why their effects aren’t immediately visible in productivity stats. In their book, published in 2014, Brynjolfsson and McAfee reference Paul David, an economic historian. After digging through records of American factories when they were first electrified, around the turn of the 20th century, David found it took several decades for plants to retool to optimize the new source of power. Writes Brynjolfsson: “Only after 30 years—long enough for the original managers to retire and be replaced by a new generation—did factory layouts change.” In other words, it wasn’t electricity but the eventual shift to assembly-line production that eventually kicked off a productivity boom.

Brynjolfsson says companies are in the early stages of figuring out how to retool their processes to take advantage of digital tools such as big data and machine learning. He also says our current method of measuring gross domestic output, and by extension productivity, does a poor job of capturing the value of free goods. “If you’re giving an app away for free that does something you used to pay for, then it’s going to initially make GDP smaller,” Brynjolfsson says.

Hal Sirkin, a senior partner at Boston Consulting Group, points out that robots currently perform only about 10 percent of manufacturing tasks. “We project that over the next 10 years, that might increase up to 20 or 25 percent. So there is a long way to go on that productivity curve.” Sirkin and his colleagues at BCG forecast that by 2025, wide-scale adoption of advanced robots will increase productivity as much as 30 percent in some industries, including machinery and appliance manufacturing, and lower total labor costs 18 percent.

Whatever you believe about technology’s role in productivity, there’s broad consensus that the outlook for unskilled workers isn’t good. In a speech this November, Bank of England Chief Economist Andy Haldane said he and his staff had modeled the effects of automation on the U.S. and U.K. labor markets and concluded that 80 million jobs in the U.S. and 15 million in the U.K. were at risk.

A BCG report from September 2015 that examined the impact advanced technology could have on Germany’s manufacturing sector concluded that if 50 percent of companies adopted new tools such as autonomous robots and 3D printing by 2025, industrywide revenue could rise 1 percent, leading to an additional 350,000 jobs. If revenue were to rise only 0.5 percent, however, the result would be a net loss of 40,000 jobs. “It’s a real possibility that if we do nothing, that inequality can get worse and more people end up getting left behind,” Brynjolfsson says. “But it’s not inevitable, and it comes down to a set of policy decisions we make. If through technology we can create more and more wealth for less and less work, then shame on us if that’s a bad thing.”

The bottom line: Tech boosters say automation and other digital innovations haven’t penetrated enough to show up in economic data.

Robots to the Rescue in Davos
    Before it's here, it's on the Bloomberg Terminal.