The Technology Trap

by Carl Benedikt Frey

  The Port Clinton story is regrettably a typically American one. In the postwar years there were many places like Port Clinton, whose modest prosperity was built on the foundation of manufacturing that provided stable, thriving communities and opportunities for those starting at the bottom. But the employment prospects of blue-collar Americans are no longer what they were in the 1950s and 1960s, and neither are the communities in which they live. In Coming Apart, political scientist Charles Murray shows that ordinary middle-class Americans who could have prospered in Port Clinton in the postwar years have become increasingly detached from the rest of American society more broadly. To illustrate this, he creates a statistical construct he calls Fishtown—after the predominantly white, blue-collar community of the same name in Philadelphia, Pennsylvania—drawing upon demographic data from the Current Population Survey. Those assigned to Fishtown are all white citizens who did not go to college. If they have a job, they work in blue-collar occupations, provide in-person services, or are employed in low-income clerical jobs. What these people have in common is that their skills are insufficient for them to compete successfully in the new economy. As Murray rightly points out, “The higher-tech the economy, the more it relies on people who can improve and exploit the technology, which creates many openings for people whose main asset is their exceptional cognitive ability.”7 Thus, to paint the growing divide between the fortunes of America’s cognitive elite and the misfortunes of the white working class, he creates another statistical construct that he calls Belmont, after the upper-middle-class suburb near Boston, Massachusetts. Those assigned to Belmont have much in common with Robert Reich’s symbolic analysts, who have been reaping the benefits of computerization. They hold at least a bachelor’s degree, work in tech jobs or skilled professions, and enjoy a secure and affluent lifestyle.

  * * *

  Murray then examines in more detail what has happened to the citizens of Fishtown since the postwar boom years. One worrying albeit unsurprising trend is that fewer people work. In 1960, there was not much difference between Belmont and Fishtown in terms of working habits: in 90 percent of Belmont households, there was at least one adult who worked forty hours per week or more, and the same was true of 81 percent of Fishtown households. But by 2010, this gap had increased dramatically. In 87 percent of Belmont households, at least one adult still put in a minimum of a forty-hour work week. But in Fishtown dramatic changes in the likelihood of an adult being in the workforce had occurred, with only 53 percent of households still containing at least one working person.

  As joblessness in Fishtown grew, so did crime. Inmate surveys carried out by the federal government in the period 1974–2004 show that about 80 percent of whites in state and federal prisons came from Fishtown and less than 2 percent were from Belmont.8 Hence, the upsurge in crime and imprisonment was concentrated in one segment of the white population—the working class. And just as Wilson found that joblessness was a cause of the growing share of single-parent households among blacks, there was a fall in marriage rates among blue-collar whites.9 As late as 1970, only 6 percent of births among non-college-educated white women were out of wedlock. Four decades on, 44 percent of births were. Only a third of Fishtown’s children now grow up in families that include both biological parents. This matters, because children in single-parent households have much worse chances later in life. Indeed, the economist Raj Chetty and collaborators found that the strongest predictors of upward mobility are measures of family structure. The greater the percentage of single parents in an area, the lower the chances of their children moving up.10

  Murray does not see the computer revolution as the cause of Fishtown’s misfortunes—instead, he sees joblessness as a consequence of a deteriorating work ethic and welfare dependency. Scholars still debate the relative importance of economic and cultural factors in shaping the patterns of joblessness, crime, and marriage rates, but it seems most reasonable to emphasize the importance of both. However, there can be no doubt that economic dislocation can explain a large part of such patterns. Those assigned to Fishtown clearly belong to groups that have seen diminishing opportunity due to trade and technological change. And many of Fishtown’s social ills can be directly linked to outcomes in the labor market. Crime rates, for example, are related to the expected costs and benefits of illegal activity.11 If workers’ expected earnings in the labor market fall, so does the opportunity cost of spending time in prison. Thus, unsurprisingly, people have historically been more likely to become involved in illegal pursuits when the relative payoff of criminal activity has increased. For example, as Britain industrialized and workers saw their skills made redundant by the mechanized factory in the nineteenth century, older workers—especially those in artisan jobs—began to commit more crimes for economic gain.12

  While unemployment is often cyclical and short-lived, the decline in the wages of the unskilled due to trade and technological change has been going on for several decades and will likely have more long-lasting effects on criminal activity than short-term unemployment does. Indeed, when the economists Eric Gould, Bruce Weinberg, and David Mustard studied the relationship between vanishing opportunity in the labor market and crime, they found the former to have caused the latter. Both unemployment and falling wages affected crime rates among unskilled men. Over the period the authors analyzed (1979–97), the wages of unskilled men fell by 20 percent, while the property crimes rose by 21 percent.13

  Other misfortunes have also afflicted the people of Fishtown, which are less obviously but nonetheless directly linked to the state of the labor market. While marriage has become less common across all spectra of society, the reason it has become so much more uncommon in Fishtown is that work has disappeared. Among the white population, the skilled have consistently seen the highest percentage of married men in America since the 1880s.14 The difference in marriage rates between people in skilled occupations and blue-collar Americans narrowed until the end of the manufacturing boom years, when it experienced a reversal. The narrowing happened as the position of blue-collar workers in the labor market strengthened. But in today’s higher-tech economy, blue-collar workers are becoming less likely to marry as their position in the labor market has gotten weaker. The economists David Autor, David Dorn, and Gordon Hanson have indeed found that the disappearance of blue-collar jobs disrupted marriage markets and reduced marriage rates. Deteriorating labor market prospects made men less marriageable. One reason had to do with joblessness and the declining economic stature of blue-collar men. But more alarmingly, vanishing factory jobs also increased the mortality gap between young men and women. As work disappeared, young men became more likely to experience an early death.15

  Their findings are consistent with those of other studies. Workers who were displaced due to plant downsizings in Pennsylvania during the recession of the early 1980s suffered annual earnings losses averaging 25 percent even six years after they lost their jobs, and immediate losses of more than 40 percent.16 But layoffs did not just affect workers’ earnings. They also brought a significantly higher risk of death. When the economists Daniel Sullivan and Till von Wachter went back and traced the fates of the workers who had been displaced by the downsizings, they found that the ones who saw their jobs disappear experienced a 50–100 percent increase in short-term mortality rates after being laid off.17 Even if the effect tailed off over time, a middle-aged man saw his life expectancy drop by one and a half years, an effect comparable to that of being forty pounds overweight at the age of forty. So in 2015, when two Princeton economists, Anne Case and Angus Deaton, winner of the Nobel Prize in Economics, shockingly found that annual death rates among middle-aged whites had risen since the turn of a century after decades of improvements, they naturally suggested that the reversal might reflect the long-standing process of diminishing opportunity in the labor market for working-class whites, whose departure from the middle of society’s spectra has come with so much distress.18 Rising mortality, they found, was not caused by ty
pical killers like heart disease and diabetes but by suicide and substance abuse.19

  Reports on subjective well-being, it is true, have consistently shown that people who experience unemployment are significantly less happy, even when a wide range of factors (including income and education) are controlled for.20 Men fare the worst mentally from unemployment, especially if it occurs in their prime years.21 One widely cited study even found that “joblessness depressed well-being more than any other single characteristic, including important negative ones such as divorce and separation.”22 But while there is compelling evidence to suggest that health and well-being are closely related to labor market outcomes, to what extent the loss of jobs due to technology and trade can account for the recent upsurge in the “deaths of despair” documented by Case and Deaton remains an open question. The growing misuse of and addiction to opioids has turned into a serious national crisis that affects public health and social welfare. America’s opioid crisis is certainly part of the story, but part of it might also be the consequence of rising joblessness. What is beyond question is that disappearing middle-income jobs have caused much material and emotional suffering, which has had a devastating impact on a broad swath of the middle classes.

  The Geography of New Jobs

  The drifting apart of American society is about more than unequal gains. Much more worrying than rising income inequality is the fact that large groups in the labor market have been left worse off economically and in terms of subjective well-being. Their reality has also become harder to comprehend because they have become increasingly segregated from the rest of society. Though Fishtown and Belmont are statistical constructs, they speak to a well-documented rise in geographic polarization. When the eminent sociologist Douglas Massey, together with Jonathan Rothwell and Thurston Domina, recently examined patterns of segregation in America, they found that a kind of incipient cognitive class apartheid had worked itself through the country in the last quarter of the twentieth century. College-educated parents and their children had become increasingly detached from the realities facing those families who have seen work ebb. The growing split between the college educated and the rest is not just one of neighborhood segregation.23 A broader shift has also taken place between cities, which is linked to the geography of new jobs.

  In the 1980s and 1990s, it was believed that the exact opposite would happen. With the advent of the World Wide Web, email, and cell phones, pundits proclaimed that location would soon become irrelevant and the curse of geography a distant memory.24 Futurists like Alvin Toffler even predicted that the death of distance would eventually render the city obsolete.25 And in 2005, the cover of the first edition of Thomas Friedman’s best-selling The World Is Flat pictured a world in which geographic divisions were history.26 Information technology, these authors declared, was making face-to-face interactions unnecessary, so that the time when companies and workers had to cluster in expensive places like Manhattan or Silicon Valley would soon be over. But the truth is that even with modern computers, many complex interactions remain too subtle to be performed via technology. As Harvard University’s distinguished economist Edward Glaeser pointed out at the time, digital and in-person communication are best seen as complements rather than substitutes for each other.27 More efficient information technology has made it possible to maintain a greater number of relationships, which increases the number of in-person contacts. Though the computer revolution rendered New York’s advantages as a manufacturing city obsolete, it amplified its competitive edge in innovation. Cities specializing in knowledge work (that is, developing and exporting ideas) became more productive.

  Place still matters because of “agglomeration economies,” which derive from the value of proximity. Workers want to be close to jobs. Companies want to have access to talent and be close to customers. Parents want to live near good schools. And the elderly may prefer places with good health care and a pleasant climate. Agglomeration, in short, comes down to the desire to reduce the costs of moving goods, people, and ideas.28 Of course, the cost of transporting goods has become a much less important factor in where companies chose to locate, simply because shipping has become so much cheaper. One reason why industrial cities like Detroit began to decline before the age of automation, when manufacturing employment was still expanding, is that production began to move away from the Great Lakes to right-to-work states in the Sunbelt, where union security agreements between companies and labor unions are prohibited.29 Such locational freedom, however, may have reduced the desire to transport the smart people and ideas that have become so valuable in the higher-tech economy. And indeed, that is what has happened.

  In The New Geography of Jobs, the economist Enrico Moretti tells an intriguing story of two places in California: Menlo Park and Visalia. The story begins in 1969, with a young engineer turning down a job offer at Hewlett-Packard in Menlo Park (in the heart of Silicon Valley) to move to the midsize town of Visalia, three hours’ drive away. At the time, many professionals were leaving cities for smaller communities, which were considered better places for family life. At the time, both places in California had prospering middle classes, similar rates of crime, and comparable quality of schools. And while incomes in Menlo Park were higher on average, America was on an equalizing path.

  Yet today, Menlo Park and Visalia are in different universes. As Silicon Valley has grown to become the world’s hub for innovation, Visalia has become a backwater. It has the second lowest share of college-educated workers in America, and its crime rates are high and trending upward while its relative earnings are in decline.30 And these are not isolated examples:

  [They reflect] a broader national trend. America’s new economic map shows growing differences, not just between people but between communities. A handful of cities with the “right” industries and a solid base of human capital keep attracting good employers and offering high wages, while those at the other extreme, cities with the “wrong” industries and a limited human capital base, are stuck with dead-end jobs and low average wages. This divide—I will call it the Great Divergence—has its origins in the 1980s when American cities started to be increasingly defined by their residents’ levels of education.… At the same time that American communities are desegregating racially, they are becoming more segregated in terms of schooling and earnings.31

  This trend began with the age of computers. It is true that some professional services, like accounting, can now be delivered electronically from a distance. But new jobs, spawned by computer technologies, are highly concentrated, suggesting that place has become more significant as production has become more skill intensive. This is underlined by the dramatic shift in the geography of new job creation beginning in the 1980s. Occupational classifications, which are updated every decade, allow us to identify new jobs that did not exist a decade earlier. My work with Thor Berger, an economic historian, shows that before the computer revolution, when some emerging occupations were still routine, new jobs didn’t just emerge in skilled cities. But as a wide range of computer-related occupations—like those of computer programmers, software engineers, and database administrators—became more plentiful in the 1980s, the comparative advantage in new job creation firmly shifted toward cities initially specializing in knowledge work (figure 15).32 We found in a follow-up study that data on the location of new industries, as opposed to new occupations, reveal a similar pattern. New industries that appeared in official statistical classifications for the first time in the 2000s primarily relate to digital technologies, such as online auctions, web design, and video and audio streaming. Ironically, it is precisely the technologies that futurists once believed would flatten the world that have made it more uneven: digital industries have overwhelmingly clustered in cities with skilled populations.33

  FIGURE 15: Knowledge Work and New Job Creation in U.S. Cities, 1970–2000

  Source: T. Berger and C. B. Frey, 2016, “Did the Computer Revolution Shift the Fortunes of U.S. Cities? Technology Shocks a
nd the Geography of New Jobs,” Regional Science and Urban Economics 57 (March): 38–45; J. Lin, 2011, “Technological Adaptation, Cities, and New Work,” Review of Economics and Statistics 93 (2): 554–74.

  Note: These figures show the percentage share of each city’s workers that were employed in jobs that did not exist by the beginning of each respective decade against the initial share of “knowledge workers” in occupations that involve abstract tasks across 321 American cities.

  The location decisions of technology companies, which are at the forefront of digital technology, provide the best evidence of the value of in-person communication: “The fact that Silicon Valley is now the quintessential example of industrial agglomeration suggests that the most cutting-edge technology encourages, rather than eliminates, the need for geographic proximity.”34 The past two decades have supported that view, as geographic clusters like Silicon Valley remain strong, despite the abundance of long-range electronic communication tools. In fact, geography has become more important as new jobs have become more skill intensive. Indeed, for most of the 1900s places with lower average incomes were catching up with richer cities and regions. One of the most widely cited studies in economics is a paper by Robert Barro and Xavier Sala-i-Martin that shows persistent and speedy income convergence across American regions over the century preceding the computer revolution.35 Not just in America but also across the Atlantic, the great convergence within countries was pervasive and persistent throughout the postwar decades. But income convergence came to a halt in the 1980s, as cognitive segregation became more widespread. In a paper titled “Why Has Regional Income Convergence in the U.S. Declined?,” the Harvard University economists Peter Ganong and Daniel Shoag argue that historically, income convergence was driven by people migrating from poorer regions to wealthier ones. The steady influx of workers held down wage growth in rich places and helped incomes rise in poor regions as people left. But the clustering of new jobs in skilled cities in combination with stricter land-use regulations disrupted this trend: the rising costs of living in booming cities meant that migration ceased to be an option for the unskilled, while highly skilled workers continued to migrate.36