Is the increasing automation of our economy a threat to American wages and jobs? Should the American worker fear the rise of the robots? No, not really.
Eighty years ago, John Maynard Keynes warned that society faced “a new disease” of “technological unemployment” in which the “means of economizing the use of labor [were] outrunning the pace at which we can find new uses for labor.” Much more recently, Michael Strain of the American Enterprise Institute wrote about how “robot workers could tear America’s social fabric.” Strain worries that machines could eliminate the livelihoods of millions of less-skilled workers.
Computers have certainly automated many tasks. From travel to banking to manufacturing to retail, machines now perform formerly human tasks quickly and reliably. Technology has eliminated countless jobs in the U.S. and around the world. Even Foxconn, famous for its vast iPhone-assembly lines in Taiwan, plans to install a million robots.
But almost as quickly as technology has eliminated some jobs, it has created new ones. Like developing smartphone apps. Or shuttling Uber passengers. Or moving inventory in Amazon warehouses. Contrary to Keynes’s prediction of 15-hour workweeks, the economy has always found new uses for displaced workers.
Indeed, automation drives growth in living standards. In order for the average American to consume more, the average worker must produce more. Automation enables businesses to make more goods with less labor, which means more output and higher living standards.
A construction worker who can operate a backhoe will make much more than one using only a shovel. An economy with backhoes will also be able to build a lot more.
In a world with more automation, not only will work still exist, it’ll be safer. Computers have automated many of the more-demanding manual-labor jobs in the economy, and workplace injuries and deaths have fallen steadily as machines took over these more physically dangerous tasks. Labor-saving technology benefits society.
Of course some people will wind up worse off than before. Some whose jobs get automated will have difficulty finding work that pays as much. And higher demand for non-routine skills will put less-skilled workers at a relative disadvantage. But the vast majority of workers will almost certainly come out ahead.
Strain worries that will not happen. He fears large numbers of those whose jobs get automated will wind up unemployable and on the margins of society. He points to two pieces of evidence: slower male wage growth since the 1970s and dropping marriage rates among less-educated men – the latter a sign, he argues, of poor job opportunities.
But it’s hard to see automation driving either trend. If machines were automating good jobs, this ought to hurt both men and women. But women have done quite well. Female employment and earnings have risen considerably since the late 1970s. As a result, overall earnings have risen considerably too.
Congressional Budget Office data shows that, after adjusting for inflation, the median market income of nonelderly households has risen by over a fifth since 1979 (on top of generally lower taxes since then too). The compensation for workers in the bottom quintile jumped over 40 percent during that period. If automation has undercut living standards, it certainly does not appear in the data.
And male wage growth has been slower, but it looks more like previously protected workers are facing human competition than robots are eating their jobs. Social barriers that prevented women from working outside the home fell away in the 1970s. Women entered the paid workforce, including once exclusively male occupations, in large numbers. As glass ceiling after glass ceiling shattered, men faced much more competition.
Simple supply and demand suggests this would raise female earnings while restraining male wages — which is exactly what happened. While the median man’s wages have grown only slightly since the late 1970s, the median woman now makes over half-again more.
Moreover, it seems hard to blame economic changes for male social problems. These problems show up in kindergarten. Research finds girls are “substantially more attentive, better behaved, more sensitive, more persistent, more flexible and more independent than boys” before first grade. These behavior gaps grow throughout middle and high school.
Far more women than men now graduate from high school and go onto college. This hurts many men’s economic prospects and reduces their marriageability. Researchers have theories about what causes these male behavioral problems, but automation does not appear on that list.
None of this makes mechanization unimportant. Economists have extensively studied the labor markets’ shift toward non-routine jobs immune to automation. Automation has made skills and education increasingly important for those trying to get ahead. Policy reforms that improve skills — such as expanding access to charter schools — have become more urgent than ever before.
But automation will not break down society. It will raise wages and living standards while reducing the drudgery of work. Don’t fear the robots.
— James Sherk is the senior policy analyst in labor economics at the Heritage Foundation’s Center for Data Analysis.