It’s rare that a day goes by when there is not some major story of workers losing jobs to self-driving cars and trucks or robots stocking supermarket store shelves or dishing out fast food hamburgers. The specifics may differ, but the story is the same: New technology will lead to mass unemployment. The frequency of these stories is truly striking for the simple reason that it’s so obviously not true.
The story of mass displacement of workers by robots is a story of rapid productivity growth. Robots are supposed to be doing the work formerly done by people. This means that we should be seeing far more output for each hour of human labor. This is something we can easily check, since the Bureau of Labor Statistics (BLS) puts out data on productivity growth every quarter.
Rather than going through the roof as the robot story would imply, productivity growth has fallen through the floor. It’s averaged just 1.2 percent annually in the last 10 years and 0.6 percent in the last five years. By comparison, productivity growth averaged 3.0 percent in both the decade from 1995 to 2005 and the long Golden Age from 1947 to 1973.
It’s possible to point to many great advances in technology. It’s also possible in future years that innovations like driverless cars will lead to mass displacement of workers, but to date the data are very clear: This mass displacement is not happening. Insofar as people are seeing job opportunities disappear, it is due to factors other than robots.
Insofar as people are seeing job opportunities disappear, it is due to factors other than robots.
It actually would be a good thing if we did see more rapid productivity growth. The period from 1947 to 1973 was a period of low unemployment and rapid wage growth. Workers were able to capture the benefits of rapid productivity growth in higher pay.
There was a similar story with at least the first half of the 1995 to 2005 productivity uptick. In the late 1990s, unemployment rates fell to levels that many economists did not think would be possible without triggering spiraling inflation. (It didn’t.) In addition, workers at the middle and the bottom of the wage ladder saw sustained real wage growth for the only time in the last four decades. This period of prosperity ended when the stock market crash brought on the 2001 recession, from which the labor market recovered slowly, but few economists would see the productivity boom of this decade as being a negative from the standpoint of workers.
If the productivity story is unambiguous – there is no mass displacement due to robots – there are some who argue that robots are still leading to a redistribution from workers to the people who own the robots. In other words, there is something about the nature of robot technology that affects the labor market differently than other labor saving technologies.
While this argument is highly dubious on its face, it is worth looking more closely at what it implies. “Owning” a robot is not a technological relationship. Robots would not be expensive because of the materials and labor that go into assembling them.
If we just considered the cost of physically producing robots, they should be cheap. We should all be able to buy a robot for a few hundred dollars that would cook our food, clean our house, mow our lawns, and do all sorts of other tasks that are time-consuming, unpleasant, and often involve substantial expenses. In this case, robots should be leading to rapid increases in real wages and living standards.
However, if robots are expensive and therefore redistributing large amounts of money from ordinary workers to the people who own robots, it is because of the patent and copyright monopolies associated with building robots. But these monopolies have nothing to do with the technology; these are incentives the government gives to support innovation. In other words, the length and strength of patents and copyrights are determined by public policy.
If these protections were leading to upward redistribution it would indicate that we have made patent and copyright protection too long and/or too strong. This is especially true if we aren’t seeing much payoff in the form of higher productivity growth. Robots don’t provide an alternative to the explanations for inequality that attribute it to deliberate policy choices. On more careful examination, the robot story ends up being just one more policy based explanation like trade, the weakening of labor unions, declining minimum wages, and contractionary macroeconomic policy.
The robot story is likely attractive to many people since it appears to pin the blame for inequality on the natural development of technology. This undoubtedly explains why we hear it so frequently even though there is zero evidence to support it. Maybe if the proponents understood their own argument better they would stop repeating it.