The Technology Trap

by Carl Benedikt Frey

  Still, we must distinguish between different types of technologies. If people believe that they will eventually be made better off by technological progress, they are more likely to accept the churn. But if citizens do not see their incomes improve over several decades as their alternative job options gradually fade, they are more likely to resist the force of technology. As we have seen, during the early days of industrialization not everyone came out ahead, and quite rationally those who were negatively affected vehemently opposed the introduction of new technologies. While Engels’s pause came to an end eventually, and ordinary people were much better off in the very long run, many of those who lost their jobs to machines never saw the gains from growth. We are now living through another period of worker-replacing technical change. As Rodrik points out, “The potential benefits of ongoing discoveries and applications in robotics, biotechnology, digital technologies, and other areas are all around us and easy to see.… Many believe that the world economy may be at the cusp of another explosion in new technologies. The trouble is that the bulk of these new technologies are labor-saving.”62

  In the end, there is nothing that ensures that citizens will accept the verdict of the market, regardless of whether outcomes have been shaped by automation, globalization, or some other factor. Many citizens quite understandably do not feel sanguine about recent advances in technology either. On May 23, 2018, almost all of the Las Vegas Culinary Workers Union’s twenty-five thousand members voted to go on strike. In addition to higher pay, they demanded greater job security against robots. Chad Neanover, a cook at a Las Vegas hotel, said: “I voted yes to go on strike to ensure my job isn’t outsourced to a robot.… We know technology is coming, but workers shouldn’t be pushed out and left behind.” The secretary treasurer of the Culinary Workers Union added: “We support innovations that improve jobs, but we oppose automation when it only destroys jobs. Our industry must innovate without losing the human touch.”63

  More broadly, a 2017 survey by the Pew Research Center of 4,135 American adults shows that 85 percent would support policies to restrict workforce automation to hazardous jobs, with 47 percent favoring the idea “strongly.” Another 58 percent responded that “there should be limits on the number of jobs that businesses can replace with machines, even if those machines are better and cheaper.” Respondents were more divided on the question of who should bear the responsibility for workers who lose their jobs, with about half saying it should be the government, “even if it means raising taxes substantially.” Respondents identifying as Democrats typically favored a stronger role for government, while Republicans were more likely to think it is the responsibility of individuals. Yet 85 percent of Democrats and 86 percent of Republicans thought that automation should be limited to dangerous or unhealthy jobs. And perhaps unsurprisingly, groups in the labor market that have seen jobs ebb were more likely to favor policies to restrict automation: among the respondents with only a high school diploma or less, seven in ten said that there should be limits on the number of jobs firms can allow machines to perform, with four in ten of those with a college degree sharing that view.64

  History tells us that political elites may block technological progress if they fear it may cause political unrest.65 As discussed above, preindustrial monarchs, who held most of the political power, were worried that they might have to share it with the increasingly wealthy merchant class. And they were also alarmed by the threat from below, fearing that machines’ taking workers jobs would lead to social and political upheaval. But while such concerns persisted well into the nineteenth century, growing competition among nation-states changed the calculations of the ruling classes. As noted in chapter 3, the weakening of the guilds and growing international competition meant that the external threat of replacement became greater than the internal threat from below. Cascading competition reduced incentives to block progress, in large part because “falling behind technologically makes countries vulnerable to foreign invasion.”66

  The main reason why British governments began to side with the innovators, even if progress came at the expense of middle-income workers, was concern over Britain’s competitive position in trade. They were also aware that a strong war machine depended on having a strong economy. The external threat, in other words, was perceived to be greater than the internal threat posed by machinery rioters. As the Luddite riots swept across Britain, yet another Parliamentary committee heard petitions from cotton workers suffering as a result of mechanization. The 1812 report that came out of the hearings attests to the government’s determination to keep the technological genie out of the bottle, even though it acknowledged the distress inflicted upon the workforce. As noted, Lord Liverpool, who became prime minister in 1812, was convinced that any measures to relieve workers’ hardship would only impede their redeployment, to the detriment of British competitiveness.67

  Governments in the nineteenth century did not see technology as an unstoppable force. Rather, they had to use considerable force to make sure that the Luddites and other groups were unable to block mechanization. And the working class didn’t view mechanization as inevitable either. One attempt to bring the spread of machines to a halt was followed by another. If the Luddites had been successful, the mechanized factory would have failed to replace domestic industry, and it is quite likely that the Industrial Revolution would not have begun in Britain.

  The past few decades have not seen any slackening of competition. Just as the Industrial Revolution was partly caused by competition in trade, the drive to computerize has in part been driven by high labor costs in the West and intensifying global competition. The ascent of Japan, Korea, and more recently China has left many American companies with the choice of either moving production offshore or automating. Donald Bennett, a union leader at General Electric’s Louisville, Kentucky, factory, told the New York Times in 1984: “The automation had to be done, otherwise we would have lost the plant altogether. Some jobs have been lost for the moment, but we had to accept some changes to keep the factory here. We sure as hell didn’t want those jobs to go somewhere else.”68 Even China has recently made a drive to automate to keep low-cost competition at bay. It is now the largest market in the world for industrial robots.

  The race for world technological leadership has if anything intensified in recent years. Supercomputers are the elephants of the computer kingdom, and some readers may be surprised to learn that the fastest supercomputer is no longer on American soil. This matters because the country with the fastest supercomputers will also be at the forefront of a variety of other domains—which is why the White House Office of Science and Technology Policy has singled out supercomputers as “essential to economic competitiveness, scientific discovery, and national security.”69 In 2015, the administration of President Barack Obama created the National Strategic Computing Initiative to ensure that the American lead in supercomputing is maintained. Despite these efforts, the world’s fastest supercomputer is now in China. America is feeling the pressure, because as every American government knows, shifts in technological supremacy come with shifts in political power.

  Yet in liberal democracies with shrinking middle classes, the internal threat of political unrest is becoming ever greater. And with populism on the march, the concerns of the unskilled have become harder to ignore. Even if governments are concerned with international competition, populists may choose to promote policies to restrict automation, the way they are now clamping down on globalization. Automation does not need be seen as an inexorable fact of life. Instead, it provides opportunities and challenges that governments can seek to control politically. Restraining technological innovation, for example, is not the same as restricting some of its uses. If there are strong political preferences for job conservation, policies that favor jobs at the expense of productivity might still be implemented. One reason that you are likely to pass dozens of bookstores while taking a walk in Paris is that France recently passed a so-called anti-Amazon law, which says that onli
ne sellers cannot offer free shipping on discounted books. The law is part of a drive in France to promote “biblio-diversity” by helping independent bookstores compete.70 France decided to forgo productivity and consumer benefits in the interest of keeping jobs.

  This example is not offered as an endorsement of anti-automation policies. As history shows, labor-saving technology and rising productivity are a prerequisite for rising living standards over the long run. One reason why growth was so slow before the Industrial Revolution was precisely the resistance to technologies that threatened to render the skills of the workforce obsolete. The point is that there is nothing to ensure that technology will always be allowed to progress uninterrupted. It is perfectly possible for automation to become a political target. The twentieth century was an extraordinary period in human history in that it saw very little resistance to machines. Though political parties often try to represent the interests of particular groups over time, it is also true that as the composition of the electorate shifts and new economic and societal issues arise, as do new political agendas. Politicians are autonomous actors who seek power by mobilizing voters through continuously shifting their agendas to reflect voters’ concerns, and one such concern these days is clearly automation. Consequently, in Britain, the leader of the Labor Party, Jeremy Corbyn, has pledged to tax robots to slow down the pace of automation, which he thinks threatens workers’ jobs.71 And in South Korea, President Moon Jae-in has already downsized tax breaks on investments into robotics and automation due to employment concerns.72

  In America, Andrew Yang, who will make automation the key theme for his 2020 run for the White House, thinks it is hard to tax robots directly. Instead, he proposes a special value-added tax on companies that use automation.73 Though few candidates have made automation a central political theme, we know from recent events that populist ideas can spread quickly if they gain some traction. Trump’s promise to slap U.S. tariffs on steel and aluminum imports has earned him praise from Democrats in the Rust Belt states who seek to appeal to the electorate. For example, Senator Sherrod Brown recently told Reuters that “this welcome action is long overdue for shuttered steel plants across Ohio and steelworkers who live in fear that their jobs will be the next victims of Chinese cheating.”74

  Looking forward, as labor markets become more shielded from the impacts of trade and the replacing effects of technology become greater, the populist target could shift in the coming years. The rise of China as a major industrial power has already happened, and the jobs that can be sent abroad have already left American and European soil. They are not going to come back in larger numbers, as voters will find out eventually. Those who think that tariffs on steel will bring American jobs back would do well to visit European steel mills. In Austria, fourteen employees are needed to produce 500,000 tons of steel wire per year. As a visitor at an Austrian plant notes, “There’s barely anybody there. At most, there are three technicians monitoring the output on flatscreens.”75

  Because most Americans now work in nontradable sectors of the economy, they are also increasingly shielded from the direct impacts of trade. Michael Spence, winner of the Nobel Prize in Economics, and Sandie Hlatshwayo have shown that nontradable services might have accounted for as much as 98 percent of total U.S. employment growth in the period 1990–2008.76 But as we shall see, the rise of artificial intelligence and autonomous driving means that a large percentage of nontradable jobs are now susceptible to automation. As Obama noted when leaving office, “The next wave of economic dislocations won’t come from overseas. It will come from the relentless pace of automation that makes a lot of good, middle-class jobs obsolete.”77

  Conclusion

  The rise of the middle class was in large part a consequence of two industrial revolutions. From the mid-nineteenth century until the age of computers, technological change helped a steadily growing share of workers join the ranks of the middle class. In this regard, the computer revolution was not the continuation of a century of mechanization but the complete reversal of it. Recently, automation has cut out the jobs that were created by the spread of office and factory machines over the course of the twentieth century. The restructuring of the American economy has not worked in favor of the middle class. The experience of the decades succeeding the 1980s in many ways resembles that of the early nineteenth century, when the arrival of the mechanized factory caused a similar hollowing out of the labor market, put downward pressure on workers’ wages, and caused the labor share of income to fall—to the detriment of average people. The recent populist backlash is less puzzling when considering the reversal of fortunes experienced by the non-college-educated middle class. Blue-collar families (in particular, those who felt that they had moved up in the world) now feel that they have fallen behind. It is hard to believe that populism would be as appealing if the wages of the lower middle class were rising and well-paying employment opportunities were plentiful.

  The point is obviously not to suggest that America would have been better off by stopping the technological clock at the onset of the computer revolution. We can be thankful that the Industrial Revolution was not brought to a halt by the Luddites, and like the Industrial Revolution, the age of automation has delivered enormous benefits, especially for consumers. But—again like the Industrial Revolution—it has also fundamentally restructured the economic and social fabric—to the detriment of large groups in the labor market. Parallels to the early nineteenth century can surely be exaggerated. It is hard to believe that contemporary Americans would trade their jobs for the “dark satanic mills.” And the hardships suffered by those we regard as poor today look less harsh when compared to the material conditions of the Luddites. In 2011, the Heritage Foundation published a provocative report entitled, “Air Conditioning, Cable TV, and an Xbox: What Is Poverty in the United States Today?” The authors rightly noted that the material standards of poor Americans have greatly improved over the past century. Innovative products that once were luxuries had become common in all households: “In 2005, the typical household defined as poor by the government had a car and air conditioning. For entertainment, the household had two color televisions, cable or satellite TV, a DVD player, and a VCR. If there were children, especially boys, in the home, the family had a game system, such as an Xbox or a PlayStation. In the kitchen, the household had a refrigerator, an oven and stove, and a microwave. Other household conveniences included a clothes washer, clothes dryer, ceiling fans, a cordless phone, and a coffee maker.”78 However, even the Luddites had access to a wide range of consumer items that had been unavailable to their great grandparents:

  The quantitative study of probate records and other sources has prompted historians to conclude that the increase in consumer durable goods such as clocks, furniture, toys, books, rugs, carriages, jewelry, flatware, coffee and tea paraphernalia, paintings and other domestic decorations, peaked between 1680 and 1720. Most of these goods remained primarily within the confines of the middle class—indeed they may have been the signs that defined the middle class. Yet in the eighteenth century they kept trickling down to working-class people, if not, perhaps, to the unskilled poor, cottagers and paupers, who constituted the bottom 20 percent of the income distribution.79

  The fact that low-income households today have access to many things that weren’t available to Renaissance monarchs is indeed evidence of enormous material progress over the centuries. The capitalist achievement, as Joseph Schumpeter noted, was not to provide more silk stockings for monarchs but “bringing them within the reach of factory girls in return for steadily decreasing amounts of effort.”80 But that does not make concerns over the well-being of the shrinking middle class moot. Nor does it make moot concerns over the fact that there are many necessities that are not getting any cheaper. With inflation growing faster than some workers’ wages, health care, education, and housing are becoming less affordable to many Americans, regardless of how many television sets, microwaves, smartphones, and computers they have.
Many households that are not poor but were once firmly middle class are feeling squeezed. And one reason is precisely the cheapening of many consumer products, whose cost has been brought down by automation and moving production offshore. Most Americans are both consumers and producers. In times of worker-replacing technological change, the flip side of the cheapening of goods is that significant parts of the workforce may suffer in the labor market. That is what happened during the early stages of the Industrial Revolution and again in the age of automation. Even if we assume that the redistributive effects of automation will even out in the long run, as was the case with the mechanized factory in the late nineteenth century, and technological change ends up lifting all boats, the short run can be a lifetime for some.

  PART V

  THE FUTURE

  What of the future? … The history of computing to date shows no slackening of innovation in the fundamental computational processes or in applications of computation throughout the economy. Perhaps, aside from humans, computers and software are the ultimate general purpose technology. They are a technology that has the potential for penetrating and fundamentally changing virtually every corner of economic life. At current rates of improvement, computers are approaching the complexity and computational capacity of the human brain. Perhaps computers will prove to be the ultimate outsourcer.