scorecardresearch Skip to main content

Why aren’t robots boosting economic productivity?

Robots are making inroads in the modern workplace, but the promised benefits aren’t materializing.Leon Neal/Getty Images/File

The robot revolution is real, but the effects are doubly disappointing. Not only are they displacing workers and taking manufacturing jobs, somehow robots are failing to boost output or make the economy more productive.

When, for instance, an auto manufacturer installs cutting-edge robotic arms, two things are supposed to happen. First, the company becomes more efficient, allowing consumers everywhere to reap the benefit of less-expensive, robot-welded cars. Second, while they might lay off some assembly line workers, in theory those workers could move into high-demand jobs in other fields, like health care.

Somehow, this dynamic isn’t translating from individual businesses to the economy at large.


Despite a dramatic uptick in the number of industrial robots at work across the United States, overall productivity growth has been at a historic low, tied for the worst performance of any economic cycle since World War II.

And the people displaced by robots aren’t landing other jobs. Every newly installed industrial robot leads to as many as six lost jobs, according to a groundbreaking new paper from economists at MIT and Yale.

Explaining this two-pronged problem — robots and lost jobs, robots and lagging productivity — is an urgent concern for scholars and politicians. But whatever the cause, it seems like our increasingly robot-driven world isn’t living up to the great techno-utopian promise, namely that when machines do the work, humans will reap the rewards.

Robots are here to help

Industrial robots really are gaining ground. Their use has grown more than fivefold since the early 1990s, adding up to about 1.75 robots for every 1,000 workers.

We’re not talking Roombas here, or even the kind of manufacturing machines you might see on “How It’s Made,” which do one thing over and over with remarkable precision. Industrial robots are more advanced than that: automatically controlled, reprogrammable, and multipurpose.

Perhaps the most identifiable example would be high-tech robotic arms, which can take instructions, move fairly freely, and perform a variety of tasks, from welding to painting to riveting. Or, if you prefer, think of the stubby-but-powerful boxes that move products around Amazon’s warehouses.


Most of these industrial robots are used in manufacturing, largely in the auto industry but also across electronics, plastics and chemicals, and even food and beverages. They’ve made particular inroads across the industrial Midwest and southwest toward New Orleans.

Those in the robot industry celebrate not just the many skills of these high-tech wonders but also their potential to help — especially by freeing humans from dangerous and repetitive tasks.

Tom Ryden, executive director of the Boston-based research collaborative MassRobotics, emphasizes that part of the point is to improve working conditions, not eliminate work. “I don’t really think they’re going to replace so much as change the task of jobs, especially the dull, the dirty, and the dangerous,” Ryden says.

Robots are replacing workers

It has been hard to adjudicate this argument about whether industrial robots are helping or hurting — displacing workers, or merely allowing them to move into less dangerous, better-paying jobs. But recent economic research suggests the harm robots are causing human workers is real.

A new working paper from MIT economist Daron Acemoglu and his Yale colleague Pascual Restrepo suggests that in recent decades at least, the losses have been substantial. Looking across local labor markets, they find that when a robot comes to town, three to six nearby jobs disappear. And average wages fall, too.


The overall effect is not as dramatic as some other sources of job loss, like trade with China, but it adds up to between 360,000 and 670,000 jobs lost between 1990 and 2007.

In addition, the researchers were unable to find any positive counter-effects. Jobs lost in one area don’t migrate to another. And while some groups certainly do worse than others — men, blue-collar workers, high school grads — nobody in the local community really gains from the arrival of robots: not managers, not college grads, not even those with advanced degrees.

That flies in the face of economic expectation, insofar as the need to program and oversee robot workers should increase demand for high-skilled workers.

Robots aren’t making the economy more productive

Even if automation is eliminating jobs, that’s not necessarily damning. Job losses could be counterbalanced by declining consumer prices — or it might be that if businesses failed to automate, they might lose customers, go under, and ultimately hurt far more workers.

Jeff Burnstein puts it this way: “The real threat to jobs is when companies can’t remain competitive.” Burnstein is president of the Association for Advancing Automation, a robotics trade group, and he emphasizes that if US businesses stopped automating, they’d lose out to foreign companies that take the leap. “After automating, companies are more competitive and able to get contracts that they were never going to win otherwise.”

In the past, even those most pessimistic about the effect of robots have tended to accept the force of that argument: More robots mean a more productive economy. Where weavers might once have labored for weeks on a single rug, powerlooms meant they could produce multiple rugs in a single day. Robots are just the next step in this process, pumping out rugs without the need for any weavers.


Trouble is, while industrial robots are certainly helping some businesses, and increasing the efficiency of particular industries, they don’t seem to be making our economy more efficient. The standard metric for this is productivity — how much stuff (goods and services) we create for every hour of human work.

In recent years, productivity hasn’t been increasing very quickly, only a bit over 1 percent annually. That’s one of the slowest productivity growth rates of any period in modern US history, according to the Bureau of Labor Statistics.

Some writers have taken this as a sign that the robot revolution isn’t actually happening. If it were, they reason, productivity would have to be rising more rapidly.

But there’s a growing realization that robots and slow productivity just might go together. Ryan Avent is an economics writer with a recent book on the topic, and he maintains that there’s a fundamental problem with the idea that robots will boost productivity: People need jobs to fend off poverty. So if they’re forced off the assembly line by an industrial robot, they’ll take whatever jobs they can find. And generally, in the current economic climate, that means low pay, low productivity work.


“If someone unveiled a robot tomorrow which could do the work of 30 percent of the workforce,” he argues, “employment wouldn’t fall 30 percent.” It couldn’t — because those workers have bills to pay and families to support. Instead, they would “seek out other work, glutting HR offices and employment centres,” Avent says.

That means the net effect of even the most transformative robot might be close to zero. While robot-deploying companies could boost their own productivity, the displaced workers would cancel out the overall gains by taking replacement jobs in lower-wage, lower-efficiency sectors of the economy.

This is only the beginning

Robots are only going to get smarter. Today they can be programmed to fit car parts together; tomorrow they’ll be driving; someday, perhaps, they’ll teach classes and write complex newspaper articles.

Adjusting to these intelligent machines is one of the great long-term challenges facing modern societies. And if the questions are only beginning to come into focus, the answers remain elusive.

How, for instance, to ensure that workers and business owners share in the benefits of a robot workforce? A robot tax? Universal guaranteed income? Without some solution, you risk falling into a kind of economic paradox, well described by Harris Gruman, the Massachusetts political director for the Service Employees International Union: “If you automate everything, and displaced workers have no new source of income, who is going to buy the products?”

But these are long-term questions. There seems to be a more immediate problem: Industrial robots are making inroads in the modern workplace, but the promised benefits aren’t yet materializing.

It could be a matter of growing pains, an awkward first stage before robots help to create a virtuous circle of heightened productivity and broadly shared benefits. Since the dawn of the industrial revolution, skeptics have warned about the destructive consequences of machine labor, yet somehow advanced economies have adapted to every new innovation, and it’s led to rising levels of material well-being.

Then again, maybe it will be different this time. Maybe we’re seeing early warning signs that the robot revolution won’t be as kind to humans as we might hope.

Evan Horowitz digs through data to find information that illuminates the policy issues facing Massachusetts and the United States. He can be reached at Follow him on Twitter @GlobeHorowitz.