Carl Benedikt Frey, The Technology Trap: Capital, Labor, and Power in the Age of Automation (Princeton, NJ: Princeton University Press, 2019), 480 pp., $29.95.

DEMOCRATIC PRESIDENTIAL candidate Andrew Yang has declared, “The automation of our jobs is the central challenge facing us today.” Yang’s message, echoed by another candidate, South Bend mayor Pete Buttigieg, won’t win him the nomination, but it is backed up by several social scientists including Massachusetts Institute of Technology’s (MIT) Erik Brynjolfsson and Andrew McAfee and Oxford researchers Carl Benedikt Frey and Michael A. Osborne. In 2013, Frey and Osborne predicted that in “perhaps a decade or two … 47 percent of total U.S. employment” would be at “high risk” of being automated. That could portend what futurist Martin Ford has called a “jobless future” and would call for drastic measures to prevent a social and political cataclysm.

Now Frey has written a long book, The Technology Trap: Capital, Labor, and Power in the Age of Automation, putting his findings in historical context. Frey argues that automation, or what he calls the third industrial revolution, is not only putting jobs at risk, but is the principal source of growing inequality within the American economy. The failure to meet this challenge, Frey warns, is fueling populist and white identity politics, most evident in the 2016 election of Donald Trump.

FREY’S BOOK is about a third longer than it needs to be. He and his publisher were, perhaps, beguiled by the commercial success of Thomas Piketty’s weighty Capital in the Twenty-First Century. Frey’s book is highly repetitious. And before getting to the heart of the argument, which is the difference between the first, second and third industrial revolutions, you have to wade through chapters about Neolithic and preindustrial technology. But the heart of the argument is interesting and worth pondering.

According to Frey, the West has experienced three industrial revolutions that have been impelled by technology. The first, dating from the late eighteenth century, was driven by the steam engine, the railroad and the cotton gin; the second, which extends through the first six decades of the twentieth century, by electricity and the internal combustion engine; and the third, which begins sometime in the 1960s, and is still going on, by computer technology and, most recently, artificial intelligence. Each of these revolutions has had different effects on employment and equality, depending on the kind of technology that was introduced.

The effect has depended on whether the technology was “enabling” or “replacing”—a distinction that is common among social scientists who write about automation. An enabling technology increases the productivity of existing workers without eliminating their jobs. A good example would be how the typewriter increased the power of a clerk without eliminating the need for clerks, or how computer design increased the productivity of architects without imperiling their jobs. But the ATM replaced and eliminated many bank tellers. Robots, combined with industrial reorganization, have replaced assembly line workers. And so on.

According to Frey, the first industrial revolution was dominated by “replacing” technology. Weavers and other artisans were replaced by simple machines that could often be operated by children. Some of these former artisans became low-wage farm laborers, while others were unemployed. Overall, wages and labor’s share of national income plummeted. Economic historians call this period the “Engels’ Pause”—a reference to Friedrich Engels’ classic The Condition of the Working Class in England in 1844, which documented the immiseration of the peasantry and working class under the new technology. Marx’s socialist politics was rooted in this first industrial revolution: it assumed a rebellious working class facing unremitting downward pressure on wages.

The second industrial revolution, Frey argues, was dominated by enabling technology. The key was the rise of the electricity-powered assembly line, the gasoline-powered engine and the new electric office. Productivity rose rapidly, but wages kept pace, and the gap between the wealthy and everyone else actually shrank. The third revolution has taken place in two stages. In the first, which featured robots, many mid-wage, routine industrial jobs disappeared—among those were the assembly line jobs created in the earlier revolution. In the next phase, based on artificial intelligence, many lower-skilled service jobs will disappear. These would include office and administrative support, sales, food preparation and serving, and transportation. Frey thinks the development of autonomous vehicles will soon have a devastating effect on truck drivers, who are the largest single occupational group in many states.

Factory workers who lost their jobs during the first phase of the third revolution (circa 1970–2010) often had to take lower-paid service sector jobs. The same thing will happen, Frey predicts, with workers who lose their jobs in the phase characterized by artificial intelligence. Frey’s prediction is dire. He writes, 

A truck driver in the Midwest is not likely to become a software engineer in Silicon Valley. He might take up work as a janitor. Or he might find work in grounds maintenance, keeping parks, houses, and businesses attractive … If he became a janitor he would trade a $41,340 job (2016 annual median income) for a $24,190 job. If he manages to become a ground maintenance worker, he would make $26,830 per year. Or he might get a job as a social care worker, earning $46,890 per year. But that would require him to get a college degree.

In this way, the third industrial revolution would resemble the first: it would render a mass of workers obsolete and depress overall wages. And, like the first revolution, the third could precipitate a revolt from the bottom—led initially by right-wing populists like Donald Trump or France’s Marine Le Pen. “The robot revolution is largely a Rust Belt phenomenon, and this is also where Trump made the greatest gains for the Republican Party,” Frey observes.

SOME OF Frey’s analysis of the political implications of the third revolution seems overdrawn: he claims that the research he did with two other social scientists demonstrated that if the number of robots had not increased since 2012 in Michigan, Pennsylvania and Wisconsin, Hillary Clinton would have won these states and the 2016 election. Overall, however, Frey’s analysis of the Trump vote in the Midwest is pretty astute. Some liberal commentators have attributed Trump’s votes entirely to white racism and identity politics. Frey locates it instead in a broader problem of identity created by “fading opportunity in the labor market.”

Frey argues that “industrial male workers had to find ways of taking pride in monotonous toil on a factory’s assembly line.” Citing sociologist Michèle Lamont, Frey writes that their solution was to construct an identity as a “disciplined self.” He concludes, “In Rust Belt cities and townships, where joblessness is now widespread, the ‘disciplined self’ identity has become harder to maintain, making dormant grievances come alive.” These include the cultural resentments that liberals have focused on in explaining Trump’s votes. In my own book, The Nationalist Revival, I similarly described the threat to workers’ “way of life” from a decaying industrial base, which carried with it the destruction of neighborhoods, bars, churches, union halls and of the expectation that one’s children could enjoy the same lifetime employment.

Frey’s analysis of working-class discontent also leads him to dismiss the condescending solution favored by some wealthy Democrats for economic inequality. Yang, former Facebook publicist Chris Hughes and other one-percenters have argued for a universal basic income that would provide the equivalent of a supplementary welfare payment—from $500 to $1000 a month, even to those unable or unwilling to find work.

These eupeptic plans run afoul, Frey argues, of the average Americans’ desire to earn a living through work and aversion to those who might not share this legacy of the Protestant Ethic. He quotes his Oxford colleague, Ian Goldin, who contends that “Individuals gain not only income, but meaning, status, skills, networks and friendships through work. Delinking income and work, while rewarding people for staying at home, is what lies behind social decay.” Exactly right.

In a closing section, Frey enumerates his own proposals for dealing with job displacement owing to the third industrial revolution: 

Addressing the social costs of automation, will require major reforms in education, providing relocation vouchers to help people move, reducing barriers to switching jobs, getting rid of zoning restrictions that spur social and economic divisions, boosting the incomes of low-income households through tax credits, providing wage insurance for people who lose their jobs to machines, and investing in early childhood education.

While by no means novel, these proposals make perfect sense.

FREY COVERS a lot of ground in his book, and I won’t pretend to match his erudition. But I want to raise a few questions about his historical account of the industrial revolutions and about the overweening role that he assigns to automation in explaining economic equality and inequality.

First, the history of technology and jobs: I don’t think the distinction between replacing and enabling technology fully accounts for the difference between the revolutions and their effect on jobs and economic equality.

Take the difference between the second and third revolutions. Frey acknowledges in passing that, during the second revolution, technology did dramatically replace employment, especially in agriculture. In 1850, according to some estimates, 64 percent of the country’s workers worked on farms; by 1929, due in large part to the introduction of reapers, tractors and other machinery, it was down to 18.3 percent. It is now below 2 percent, even though total production has continued to increase. During the same period, a host of crafts occupations were also replaced by assembly-line production.

What made the second industrial revolution such a boon to employment was the introduction of vast new industries. These included automobiles, aircraft, electric machinery, telephones and household appliances. Citing the historian David L. Lewis, Frey notes that by 1986, the auto-industrial complex, which consists of manufacturing, services and construction, employed one out of six Americans. Many of the workers or the children of the workers who were replaced by machinery on the farm or in crafts went to work in these industries.

Now compare the third industrial revolution, which has also produced vast new industries in software, entertainment, robotics, biochemistry, personal computing and telecommunications. Some or most of these new industries are not as labor-intensive as the industries of the second industrial revolution. But there is an added factor: in the United States, many of these industries do much of their manufacturing overseas or south of the Border. MIT economist Daron Acemoglu has said, 

Many of the important products that we have created over the last several decades, over the last two decades in particular, have not added much to the bottom line employment figure. Think of Apple’s iPods and iPads and iPhones; they are amazing innovations and consumers have absolutely rewarded the company by purchasing billions of them … So, there we have an intersection of new technologies that have a very heavy design component and the division of labor can be very finely broken down, and the labor-intensive parts of those products can be manufactured abroad. So, there is a sort of a parallel process to automation that’s increasing efficiency but it’s not really adding to the bottom line employment figure.

There is an additional factor that the Manhattan Institute’s Oren Cass cites in his analysis of the effect of automation on job growth. From 1947 to 1972, as the second industrial revolution was reaching its climax, manufacturing productivity grew 3.4 percent a year and output grew 4.2 percent a year. Yet from 2000 to 2018, productivity grew 3.1 percent a year—most likely as a result of automation—but output only rose 1.3 percent a year. In this case, the real villain in job loss was not automation, but the lag in manufacturing output. If output had grown as it had in the past, there would not have been the same overall loss in jobs due to automation.

So you have automation as a secondary factor in job loss due to the lack of a corresponding increase in output. The lack of domestic output was partly due to manufacturers shifting their jobs out of the country—in North Carolina, for instance, which Trump won, its furniture industry shifted to Mexico—and to ruinous competition from subsidized industries in Asia that knocked out American firms. In one highly-quoted study, economists David Autor, David Dorn and Gordon Hanson blamed competition from Chinese exports for the loss of up to 2.4 million jobs from 1999 to 2011. The same authors, along with Kaveh Majlesi, also argued that import competition contributed to political polarization and fueled the rise of Trump. Again, it wasn’t just automation that led to the loss of jobs and the rise of right-wing populism.

My second concern with Frey’s analysis has to do with the relationship technology has to wages and rising inequality. Frey attributes rising wages and greater equality during the second revolution primarily to wages keeping in line with rising productivity and skill levels. Child labor was replaced by more skilled adults. But I would add several factors which become relevant in judging why wages haven’t kept pace with productivity during the third revolution.

During the second revolution, there was a favorable relationship between supply and demand created by the restrictions on immigration introduced in 1920 and 1924 legislation. The rise of the labor movement in the 1930s, which was particularly strong in the new industries and which, by the 1950s, accounted for a third of the non-farm labor force, buoyed wages; and during the two world wars, American administrations took measures, including progressive taxation, that increased economic equality. (In the 1920s, by contrast, when wartime taxes were repealed and regressive changes introduced, and when a corporate offensive demolished labor unions, economic inequality rose.)

If you now look at the third industrial revolution, it becomes clear that there were other factors besides automation that have contributed to growing inequality and wage stagnation. Wage stagnation has coincided with the decline of labor unions, particularly in the private sector. This decline was not directly the result of automation, but of an employer offensive aided initially by the Reagan administration. Many plants moved south or out of the country in order to avoid unionization. Corporations used strong-arm tactics, some of which were illegal—for instance, firing organizers—to block organizing drives. When unionized labor no longer makes up a significant share in an industry’s employment, unionized firms no longer set the standard for wages—they have to adjust to their competition. By the 1990s, that had become the case in many key industries.

Wage stagnation at the bottom of the income ladder also coincided with the dramatic rise of unskilled immigration, which began after the 1965 immigration bill and accelerated in the 1990s. It depressed the wages of service and construction workers and also made it more difficult to unionize. Employers used legal and illegal immigrants to bust unions in meatpacking, agriculture, construction and janitorial services. In insisting on the primary role for automation in wage stagnation and inequality, Frey acknowledges that the decline in unionization may have played a lesser role, but he denies any role to the increase in unskilled immigration.

There is, finally, the idea that 47 percent of American jobs will be in “high risk” from automation “in a decade or two.” There is no question that some jobs will be at risk. Some, like telemarketers, tax preparers, brokerage clerks and file clerks, are already on the verge of extinction. But as Cass points out in his analysis, Frey and Osborne appear to equate the replacement of certain tasks within occupations with the replacement of the occupations themselves. An Organization for Economic Cooperation and Development study of American occupations only found 9 percent that were easily automatable. A McKinsey Global Institute study found that while at least 30 percent of activities within 60 percent of jobs could be automated, less than five percent of jobs “could be automated in their entirety.” (McKinsey also cited studies showing that by 2030, 8 to 9 percent of jobs “will be ones that barely exist today.”) Self-driving cars and trucks may be decades rather than a decade or two away from displacing drivers. And other occupations on Frey and Osborne’s list—including models, bicycle repairers, tour guides, cooks and animal breeders—do not appear like obvious candidates for machine replacement.

FREY’S BOOK is an extremely useful history of the effect of technology on jobs and income inequality, but he oversells the effects of automation. If Americans want to do something about stagnant wages and rising inequality, they have to address factors that don’t directly stem from automation. These include restoring the original function of the National Labor Relations Board, which was to facilitate rather than impede unionization and collective bargaining; repealing right-to-work laws that discourage unionization; passing comprehensive immigration reform, but also restricting the numbers of unskilled immigrants who would compete with unskilled or low-skilled workers who are already here; and removing incentives for American companies to relocate their production out of the country. Pairing these kinds of measures with those that Frey recommends could turn technology from a trap into something of a blessing.

John B. Judis is the author of The Nationalist Revival: Trade, Immigration, and the Revolt against Globalization.

Image: Reuters