Innovation Files has moved! For ITIF's quick takes, quips, and commentary on the latest in tech policy, go to

I (Want) Robot

Welding Robot

Originally published on the Huffington Post.

The new meme is that the robots are coming, and while they are not the Arnold Schwarzenegger “life terminator,” they are the jobs terminator. As MIT professors Erik Brynjolfsson and Andrew McAfee assert in their book Race Against the Machine, workers are “losing the race against the machine, a fact reflected in today’s employment statistics.” Gary Marcus added that as machines continue to get smarter, cheaper, and more effective, our options dwindle. So don’t bother polishing up that resume, rather here’s a link to the unemployment office.

These assertions parallel the largely unfounded fear of technology as a whole and often are used by interest groups and politicians to rally the masses without any clear understanding of what they are against. And as they say, it is a tale as old as time.

During the 1930s, a labor union wrote a letter to FDR proposing the following: “Remove the loading machines from the coal mines, complete all public work with man power, take the tractor off the farms, go into the various industries and remove enough labor-displacing machines to make employment for labor.” A few years later, Congress debated legislation to require the Secretary of Labor to create a list of all labor-saving devices and estimate how many people could be employed if these devices were eliminated. This is not to mention the laws that attempted to ban automobiles because they offended the horse.

When factory automation took off in the late 1950s and early 1960s, fears surrounding the employment effects of productivity increased further. They even entered pop culture with television shows, news documentaries, and reports worrying about the loss of work. One particularly telling episode of The Twilight Zone documented a dystopian world in which a manager replaces all his workers with robots, and in the final scene, the manager himself ends up being replaced by a robot. The concern caused by automation and the rise of push button factories was so great that the Congressional Joint Economic Committee held extended hearings on the matter in 1955. And in 1961, President Kennedy created an Office of Automation and Manpower in the Department of Labor, identifying, “the major domestic challenge of the sixties — to maintain full employment at a time when automation, of course, is replacing men.”

Others at the time even considered schemes whereby the United States would encourage migration of Americans to other nations as the demand for labor contracted.

At first glance this seems logical. After all, the main purpose of automation has always been to make it easier for humans to produce more with less work. But automation, whether it’s robots, computers, or machine tools, has never led to an increase in unemployment, and more importantly never will. Both history and scholarly analysis have clearly and consistently refuted the notion that increased productivity (through automation, self-service or robots) leads to higher unemployment in the moderate or long-term. In the 20th century, new technologies including tractors, disease resistant crops, and chemical fertilizers boosted agricultural productivity. While this caused a decline in agricultural employment, food became cheaper and consumers spent the money they saved from on other things (e.g., cars, appliances, entertainment) thus creating employment in other sectors. Similarly, while auto factory automation makes it possible to produce more cars with fewer workers, it also lowers the price of cars, thereby boosting demand for cars and creating employment.

Some of today’s robot alarmists will be willing to acknowledge this, but they argue that things are different now because technology is displacing jobs not only in agriculture and manufacturing, but also in the service sector. And not just in low skill jobs but high skill ones there will be no new job-generating growth sectors to employ all those who lose their jobs. Congenital pessimist Jeremy Rifkin argues when millions of retail jobs are displaced by e-commerce and a host of other service sector jobs undergo digital automation, there will be no new jobs to replace them. If we boosted productivity in the retail, banking, insurance, and other service sectors that were job generators up until now, where in the world will people find work?

This view suffers from a fundamental flaw of first order logic. Yes, automation may lead to a worker being replaced by a machine. But to understand what happens to the economy as a whole we need to ask what happens as a second order effect.

Robots free up work, which means we have more money to spend on everything from massages to motorcycles to mint juleps. Consider that the savings from a more efficient industry would flow back to the economy in one or more of the following three ways: lower prices (e.g., lower rates for consumers), higher wages for the remaining employers, or higher profits. In a competitive market, as virtually all industry markets are, most of the savings would flow back to consumers in the form of lower prices. Consumers would then use the savings to buy things (e.g., go out to dinner, buy books, go on travel). This economic activity stimulates demand that other companies (e.g., restaurants, book stores, and hotels) respond to by hiring more workers. And for anyone who says that people wouldn’t be able to find things to spend their money on, just go to a shopping mall sometime and give the average shopper $10,000 and ask them if they would lack things to spend it on.

This common sense view is borne out by economic studies. A number have found that productivity growth does have some short-term negative job impacts, but moderate- and long-term benefits. For example, Chen, Rezai, and Semmler find that while short-run productivity growth and unemployment are weakly positively correlated, in the moderate- and long-run productivity growth is strongly negatively correlated with unemployment. Likewise, in a definitive review of the studies on productivity and employment, the OECD stated:

Technology both eliminates jobs and creates jobs. Generally it destroys lower wage, lower productivity jobs, while it creates jobs that are more productive, high-skill and better paid. Historically, the income-generating effects of new technologies have proved more powerful than the labor-displacing effects: technological progress has been accompanied not only by higher output and productivity, but also by higher overall employment.

Even for those who grant that robots make the economy bigger, many still fear them. Keynsian economist Paul Krugman writes:

Smart machines may make higher G.D.P. possible, but also reduce the demand for people — including smart people. So we could be looking at a society that grows ever richer, but in which all the gains in wealth accrue to whoever owns the robots.

But there is only one way this could happen and that is if there is no competition. If one robot “owner “jacked up prices and made massive profits another robot owner would lower prices to get market share. This process would keep happening until profits were at a normal level, just as it has in the U.S. economy for the last 300 years. And to be clear, as economist Steve Rose has shown, productivity still does benefit American workers.

So, I don’t know about you, but I say bring on our robot friends. The more the better. The sooner the better.

Print Friendly, PDF & Email

  • Cirith

    The mainstream view that there will always be new jobs for humans only holds because as of now just about every human has something useful which machines still can’t do. Even low skilled workers had thinks they were superior in, since as Moravec’s Paradox states, many things we view as easy are actually much harder for computers than some things we would consider hard. It’s no wonder that previous developments never lead to permanent unemployment they were sophisitcated tools, but useless without workers. Vision, dexerty and the ability to act in an uncontrolled enviroment are hard for computers and we are just beginning to solve these problems. Proper self driving cars were only created a few year ago and it will still take time till they turn into an actial consumer product.

    However many jobs could already be automated with todays tech. I have my doubts that in a 50 years or so there will be jobs left which are low skill and impossible for a machine. In the past that may not have happened, but machines were always inferior in many ways. Looking at the past is not enough to judge what happens if the areas we are better in get fewer and fewer.

    Well not that I have anything against machines doing most jobs, it just may force us to adapt our system.