Originally posted on World Policy Institute blog on April 27, 2017 – Robot Capitalism.
Robots taking over is a staple of science fiction, including in Isaac Asimov’s classic novel, I Robot, and the films based on it. The “Marketplace” program on National Public Radio has been running a series in recent weeks that explores a more prosaic aspect of this quintessential fear: What jobs are least prone to be automated out of existence? There is another aspect of this problem that is less often considered: What use will the robot owners have for the rest of us? Perhaps it is not the takeover by robots we need to fear, but by their owners.
Economics considers three factors of production: land, labor, and capital. In economic terms, the “robot takeover” is the growing capacity of machines to replace labor. The owners of machines (capital) and land would still have income, but what about the proletarians who own only their own labor? What is their future in a robot world?
A couple years ago, best-selling author Martin Ford argued in The Rise of the Robots that the displacement of human work by machines would require something like a guaranteed income for those left without work. Yet what force will compel this? Peter Frase’s Four Futures: Life After Capitalism also considers the relationship between a guarenteed income and potential environmental degradation and property relations. Frase understands, perhaps better than Ford, that owners will not necessarily concede income to non-productive people. The mass of people may not be secure unless they own a share of the means of production themselves.
Economics textbooks have long told a happy story that though manual laborers, farmers, and factory hands are becoming obsolete, labor remains useful in services and in what former labor secretary Robert Reich called “symbolic manipulation,” referring to various jobs involving advanced mathematical and linguistic skills. However, the rapid development of artificial intelligence, symbolized by IBM’s “Watson,” will accelerate the replacement of such jobs as well. Software and machines are now so inexpensive to replicate that once a single breakthrough occurs, it can spread very quickly, displacing workers by the millions in a short time.
There are two socialist solutions that have been advocated since the 19th century. One is to continually shorten the workweek and redistribute the remaining necessary labor among more people to avoid mass unemployment. The other is to socialize ownership of the means of production so that individual or corporate owners cannot exclude most people from sharing in the bounty produced by factories and other productive facilities.
Shortening the workweek made steady progress from the later part of the 19th century into the 20th century, when a workweek of around 40 hours became standard. After World War II, some European countries went further, increasing paid vacation and shortening the workweek to something between 30 and 40 hours. Yet since the last financial crisis, the labor unions and traditional left parties seem to be in decline and disarray. Shortening the workweek further seems to be off the table for the time being.
There is even less consideration of socializing ownership of productive resources. This has never been advocated by any major party in the U.S. The left-wing parties in Europe that once championed social ownership have declined, replaced in part by right-wing nationalist groups. The decline of the traditional proletarian left comprised of labor unions and socialist parties is in part a product of automation (along with globalization).
At first, this was not so obvious. Among the most militant and effective labor unions during the mid-20th century were those in capital-intensive heavy industries like steel, automobiles, and chemicals. The idea was that, since these industries involved massive amounts of capital, often financed by debt, a strike by workers was very costly to owners, who still paid the cost of the debt incurred to buy the machines even when they were idle. Therefore, owners often found it more profitable to recognize unions and give some concessions on wages in return for labor peace. Conversely, labor-intensive industries, like apparel, fought unions doggedly since they could afford wage increases less readily.
After World War II, globalization transferred many labor-intensive jobs to poorer countries where wages were much lower and unionization rights harder to secure. At first, most capital-intensive heavy industries in developed countries remained competitive and their workers reasonably well paid. But by the end of the 20th century, even many heavy industrial jobs were moving to developing countries like China, India, and Brazil. Developed countries began to deindustrialize. Often low-paid service jobs replaced higher-paid industrial jobs, but chronic unemployment emerged as well.
Now many service roles are being automated, like call center workers, package handlers, security staff, and sales clerks. Soon drivers, computer programmers, and even teacher positions will more and more be replaced by machines. The U.S. Army is already experimenting with tutorial programs that replace many training staff. I am skeptical about the trend toward online university courses, but as these improve and become more cost competitive, they may displace many faculty.
Globalization ensures that innovations are propagated worldwide very quickly. For example, when I first went to China in 1981, the machines in textile factories were little changed from the 19th century. Steel and manufacturing companies were similar. Today, even though wages for Chinese workers are one-tenth of what is typical in developed countries, many factories in every industry are fully automated with the latest German-made, computer-controlled machinery and Japanese industrial robots. In a massive steel factory hard-hat workers are very few—low pay does not preserve their jobs. Poorer countries that have not substantially industrialized probably never will, as they likely cannot develop industries that could compete with highly automated foreign factories.
Massive worldwide unemployment and underemployment does not automatically create its own solution in a market economy. Already, many student interns in the U.S. who used to work summer jobs for at least minimum wage are applying for “jobs” that do not pay any salary at all. When labor is so abundant that its marginal value is zero, the only people who can afford to work are students whose parents are wealthy enough that they can value experience even without pay.
Globally, we may see the reemergence of a problem economist W. Arthur Lewis considered during the mid-1950s. He wrote that many poor countries have such high populations in rural areas that marginal productivity is near zero. In other words, workers could move from the countryside to the city without any drop in farm output. The abundance of such workers would keep urban wages low for a long time, until the rural labor surplus finally disappears.
Now we may have a similar problem with industrial economies. So many industrial workers will be displaced by automation that service occupations will be flooded by too many applicants, driving wages below subsistence. More are driven out as services are automated. Eventually, labor becomes so abundant that most people are economically redundant. You already see this sort of phenomenon in microcosm at tropical beach resorts where more people offer massages or trinkets to tourists than are productive and useful.
One more modern phenomenon, the rise of social media, is also displacing paid employment in entertainment industries because people can create their own content for free and entertain each other with it. The only reason this pays is that the participants are still interesting to advertisers as potential consumers, because most of them have disposable income. How long will this be true in a world increasingly polarized between the owners of intelligent productive machines and the rest of us, non-owners? If most people become unemployable, their only claim to income must come as voters in a political system that at least recognizes their value as citizens. This will tend to pit capitalism against democracy, as I argue is happening with the trend of conservatives to veer toward anarchism.