Jobs Are Under Attack, But Not by Robots
Does the last decade's slow growth in total factor productivity, which measures innovation, indicate that the dot-com revolution of 1994 to 2004 is unlikely to be repeated? How fast will the U.S. economy grow over the next 25 years?
Not as fast as Erik Brynjolfsson and Andrew McAfee assert. As techno-optimists, they believe the U.S. is at an inflection point between a past of slow technological change and a future of rapid advances. They remind us that Moore’s Law predicts endless exponential growth in the performance of computer chips, yet ignore that chips have fallen behind the predicted pace of Moore’s Law since 2005.
They also overlook that the decline in the price of information and communication technology equipment relative to performance was most rapid in the late 1990s, with little if any decline for several years. Exponential increases in computer performance will continue, but at a slower rate than in the past.
Medicine is another area in which Brynjolfsson and McAfee forecast rapid change. The most important advances in life expectancy, however, were achieved in the first half of the 20th century, when it rose at twice the rate of the second half. Infant mortality was dramatically reduced and lifespans extended by the discovery of the germ theory of disease, development of an antitoxin for diphtheria, near elimination of microbial contamination of milk and meat and near elimination of air- and water-distributed diseases through urban sanitation infrastructure.
Many of the basic tools of modern medicine were also developed between 1940 and 1980, including antibiotics, the polio vaccine and treatments for coronary heart disease and cancer. All these advances contributed to productivity growth. Medical technology hasn't ceased to advance since 1980 but rather has continued at a slower pace. Life expectancy has continued to improve at a steady rate, while the mortality rate for cardiac-related diseases has steadily declined.
Life expectancy is likely to keep improving at a rate not unlike that of the past few decades. But there are new issues. As described by Jan Vijg, a geneticist at the Albert Einstein College of Medicine, progress is advancing faster on physical than on mental disease. This raises the concern that there could be a steady rise in the burden and cost of caring for elderly Americans with dementia.
Pharmaceutical research, moreover, has reached a brick wall, with rapidly increasing costs and declining benefits, plus a decline in major drugs approved over the past decade. Drugs are being developed that will treat esoteric types of cancer at costs no medical-insurance system can afford. The current regime of drug testing in the U.S. may also be inhibiting risk-taking.
Small robots are another technology the optimists cite. Industrial robots were introduced by General Motors in 1961. By the mid-1990s, robots were welding automobile parts and replacing workers in the lung-killing environment of the automotive paint shop. Until recently, however, robots were large and expensive and needed to be caged off to keep them from smashing into humans. Smaller, more capable robots are under development due to exponential growth in computer performance and improvements in electro-mechanical design and energy storage.
Inexpensive robots suitable for use by small businesses were brought to the public's attention in 2012 by a 60 Minutes segment featuring Baxter, the $25,000 robot. Baxter's appeal is that he can be reprogrammed to do a different task each day. Other robots are mobile and can move around the factory floor. But these small robots are no different in principle from the machinery dating back to the textile looms and spindles of the late 18th century British industrial revolution.
Most work-place technologies are introduced with the intention of substituting machines for people. Yet this has been going on for two centuries. Why are there still so many jobs? Why in mid-2015 was the U.S. unemployment rate close to 5 percent instead of 20 percent or 50 percent?
MIT economist David Autor has answered this question: Machines, including futuristic robots, substitute and complement labor. Just as Baxter the robot cooperates with human workers, robots can make some people more valuable and even create new jobs, including those needed to build and program the robots.
The complementarity between robots and humans is illustrated by the cooperative ritual taking place daily in Amazon warehouses, where robots don't actually touch any of the merchandise. Rather, they lift and move shelves to the human packer, who takes the object off the shelf and performs the packing operation by hand.
The tactile skills needed to distinguish shapes, sizes and textures of merchandise are beyond the capability of current robots. Other examples of robot-human cooperation include ATMs, which have been accompanied by an increase, not a decrease, in the number of bank branches, and the barcode retail scanner, which works with but doesn't replace the check-out clerk.
The same is true for automated check-in kiosks at airports. This innovation was rolled out between 2001 and 2005 and has thinned the ranks of airport ticket-counter personnel, just as earlier airline web sites largely displaced travel agents and airline telephone agents. Yet the industry continues to employ many thousands of people as baggage handlers, flight attendants, pilots, air traffic controllers and gate agents.
Other examples abound. Goods are still placed on supermarket shelves by store employees and by the drivers of delivery trucks for beer, bread and soft drinks. Haircuts, massages and manicures are still exclusively the province of humans, as are restaurant cooks and waiters. Hotels still have front-desk personnel, and if they offer room service, it's delivered by humans rather than robots. Far from occurring overnight, the shift to robots and job-destroying artificial intelligence is occurring at glacial speed.
Surely multiple-function robots will be developed, but it will be a long and gradual process before robots outside of manufacturing and wholesaling become a significant factor in replacing human jobs in the service, transportation or construction sectors. And it is in those sectors that the slowness of productivity growth is a problem.
The ups and downs of TFP growth since the late 19th century are the result of successive industrial revolutions. The most remarkable fact about TFP performance is that rapid growth was not spread evenly over the 12-plus decades since 1890, but rather was concentrated in the middle of the 20th century. Although the digital third industrial revolution utterly changed the way Americans obtain information and communicate, it didn't extend across the full span of human life as did the second industrial revolution, with its epochal changes in food, clothing, housing, appliances, transportation, information, communication, entertainment, medicine and improvement of working conditions.
No one can foresee the future, but we can ask whether the future is more likely to resemble the 1994 to 2004 decade or the most recent decade, 2004 to 2014. The evidence is that the more rapid growth of TFP during the dot-com decade represented a temporary upsurge that is unlikely to be repeated.
How can we choose between these opposing visions? Numbers don't lie. Far from soaring toward the techno-optimists’ vision of mass unemployment, the U.S. unemployment rate has declined rapidly to 5 percent in late 2015 from 10 percent in October 2009. It seems likely to go well below 5 percent in 2016. Labor productivity, far from exploding as machines and software replace people, has been in the doldrums, rising only 0.3 percent a year in the five years ending in mid-2015, in contrast to the 2.3 percent a year in the dot-com period.
Now that the American economy has arrived back at a state of relatively full employment, it is hard to maintain the case that robots and artificial intelligence are creating a new class of the permanently unemployed.
The problem created by the computer age isn't mass unemployment but the gradual disappearance of good, steady, mid-level jobs that have been lost to robots and algorithms and also to globalization and outsourcing. Employment growth is also concentrated in routine, manual jobs that offer relatively low wages.
Slower productivity growth and low-wage jobs are leading to the unequal distribution of productivity gains. Those are the real headwinds that America faces.
This is the second of two excerpts from "The Rise and Fall of American Growth: The U.S. Standard of Living Since the Civil War," published this month by Princeton University press. Part 1 is available here.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Robert J Gordon at email@example.com
To contact the editor responsible for this story:
Paula Dwyer at firstname.lastname@example.org