“The future ain’t what it used to be,” Yogi Berra famously observed. Nowhere is that truer than regarding the future of work, particularly in cities. The economic disruption of the COVID-19 pandemic, partial de-globalization driven by Sino-American geopolitical rivalry, and the collapse of the asset bubble that triggered the decade-long Great Recession of the 2010s—all these trends have shattered the orthodox neoliberal narrative of the 2000s about the future of the American workforce, without replacing it with a new consensus.
Instead of witnessing the dawn of a new era of expanded prosperity, we are trapped in a dystopia where most workers, particularly those without elite educations and social connections, face a difficult future.
This reality is radically different from the optimistic narratives of the early 2000s. Manufacturing was out; software was in. Small towns and suburbs were yesterday; global cities were tomorrow. According to the dominant neoliberal narrative, accepted by mainstream members of both major parties and the US commentariat, globalization was an inevitable and beneficial force. The comparative advantages of the United States were the innovative knowledge economy—inventing new things but allowing them to be made elsewhere—and high-end global services such as international finance, insurance, and banking.
The optimistic globalist consensus was disseminated by journalists such as Thomas Friedman and the experts and policymakers chosen to address the rich and famous at conclaves in Davos, Switzerland, and Aspen, Colorado.1 In this view, former manufacturing workers could be retrained for better jobs in the new knowledge economy; they could “learn to code.” And if the loss of manufacturing jobs caused entire cities and regions to collapse, that was the price of progress as other regions prospered. Having learned to
code, the displaced workers could “move to opportunity,” leaving derelict Detroit to become software writers in the Bay Area. Workers who refused to upgrade their skills or “human capital” would be “left behind.”
Recognizing the New Reality
Although the shift from an industrial to a postindustrial economy was portrayed as inevitable and good, proponents of the neoliberal narrative argued that government could ease the transition—mainly by providing more STEM education for American students, many or most of whom, it was claimed, would be competing for jobs in the future global labor market with the Chinese, Europeans, Indians, and others. Yet for most Americans, particularly in metropolitan areas, this shift has been less sanguine.
At the same time, the “jobs of the future” in the knowledge economy never materialized in large numbers. Most jobs created in the past generation have been low-wage, low-benefit jobs in areas such as health care, hospitality, and retail, not well-paid jobs requiring STEM skills and college diplomas.
The trend continues. Of the top five jobs with the most job growth between 2020 and 2030, according to US Bureau of Labor Statistics projections, only one—software developers and software quality assurance analysts and testers, paying on average $110,140—can be characterized as a tech or knowledge economy job. The other four—home health and personal care aides; cooks, restaurant; fast food and counter workers; and waiters and waitresses—require no college education and pay between $20,000 and $30,000 annually, below the median income of $35,805 for a single person in 2020.2
The current boom, based on social media, search, and high-end research, does not even boost prospects for most people—even in Silicon Valley. Two left-wing scholars, Manuel Pastor and Chris Brenner, note that the area once was among the most egalitarian in the nation—a great place of opportunity for many immigrants, particularly from East Asia, who increasingly launched larger firms on their own.3 Today, they suggest, Silicon Valley has become “fragmented and divided” and characterized by “the high-tech community largely isolated from the broader region and particularly those parts . . . that are less fortunate.”4
The current tech economy, based on software, social media, and massive venture capital investments, lowers consumer prices but does not produce higher wages for most, even as it generates huge fortunes for the relative few. According to a 2018 University of California, Santa Cruz, study, nine in 10 jobs in Silicon Valley now pay less than they did 20 years ago, adjusted for inflation.5 Particularly disadvantaged are the people, most of them contractors rather than employees, who clean the offices of tech firms; make food for investors, managers, and professionals; and take care of their children.6
Some adherents of neoliberal conventional wisdom respond to these conditions by proposing to reshape benefits such as health insurance to make them portable so they would follow workers in the “new economy” as workers hopped from job to job or gig to gig. While this might help, to the extent that taxpayers replace employers as the source of adequate incomes and benefits for low-wage workers in Silicon Valley, the public is indirectly subsidizing one of the richest industries in the world.
The Limits of Neoliberalism
Industrial policy—the deliberate promotion of certain national industries considered of greater economic or strategic value than others are—was taboo during the heyday of global market euphoria between the end of the Cold War and the beginning of the Great Recession.7 The renunciation of industrial policy by the US federal government, Congress, the corporate world, and much of the media did not mean a lack of industrial policy would have minimal, or even beneficial, effects. It simply meant the landscape of American industry and the American workforce would be shaped by the strategies of major corporations, based in the US or elsewhere, and foreign countries’ trade and industrial policies.
This is precisely what happened in the US following the Cold War. America had an industrial policy—only it was not made in America. American industrial policy was made by the executives of major multinational corporations and by governments in China, India, and Mexico.
This has had dire results for many working-class Americans. As early as 2004, the economist Stephen Roach predicted:
A new global labour arbitrage—a by-product of IT-enabled globalization—is now acting as a powerful structural depressant on traditional sources of job creation in high-wage economies such as the United States. . . .
Fully 65% of the tripling of Chinese exports over the past decade—from $121 billion in 1994 to $365 billion in mid-2003— is traceable to outsourcing by Chinese subsidiaries of multinational corporations and joint ventures.8
In 2021, the Economist Intelligence Unit predicted that, as pandemic fears subsided, US-based multinationals would resume their practice of preferring to offshore production to low-wage workforces abroad.9 Multinationals’ ability to minimize labor costs by using global arbitrage to replace American workers on a mass scale marked a radical break from the past.
During the Cold War, multinationals based in the US and its European and East Asian allies and protectorates had little access to either markets or workers in Communist nations. At the same time, developing countries such as India and those of Latin America protected their domestic products and labor markets while pursuing import-substitution industrialization strategies. That changed after the Cold War ended.
From the American workforce’s point of view, the three most important nations in the new global workforce were China, India, and Mexico. China benefited the most. Its strong authoritarian state, pursuing a sophisticated mercantilist policy, compelled foreign companies to share their technologies (sometimes unwillingly) and train Chinese workers, in return for being able to use low-wage Chinese workers in global supply chains.
The Mexican government lacked the strength and competence of China’s post-Maoist regime. But East Asian, European, and US corporations built up a Mexican automobile-manufacturing sector, taking advantage of low Mexican wages and weak bargaining power by Mexican labor.
India, the third foreign labor pool that suddenly merged with the American labor market after the Cold War, specialized in low-wage call centers in India itself. It also specialized in the chain migration of Indian guest workers under H-1B, B-1, and other visas to the US, where they were employed as indentured servants by tech and finance companies dangling the prospect of employer-sponsored green cards for those who were docile and uncomplaining.
To put it another way, globalization—promoted in the 1990s as something that would allow well-paid American workers to sell products made in American factories to consumers in other countries—instead has, as its main effect, substituted foreign labor for American labor thanks to corporate strategies of “labor arbitrage” (taking advantage of differences in wages among countries or jurisdictions). The US government’s bipartisan political decision to allow US corporations to partly deindustrialize the American economy by offshoring production to save money on labor costs left most of the working class to jobs servicing the professional classes as gardeners and nannies and in restaurant and hotels.
Automation and the Future of the Urban Working Class
In 1929, John Maynard Keynes predicted possible technological unemployment because of automation.10 But to date, automation’s impact on the workforce has been limited. A 2020 study estimates that one new robot per 1,000 workers in the US reduces aggregate wages by 0.42 percent and aggregate employment by 0.2 percent.11
While automation would have eliminated many US manufacturing jobs even without offshoring and import competition, the highly robotic factories in that scenario would have been in the US, not abroad. Indeed, even in highly automated, export-oriented manufacturing sectors, employment can outstrip displacement by robots if sales to national and foreign consumers grow even faster.12
Regarding robot density—robots per 10,000 workers—the US lags behind the global leaders: South Korea, Singapore, Japan, Germany, Sweden, and Hong Kong. The US is seventh, ahead of Taiwan and China, which are rapidly catching up.13 Unlike the US, whose companies prefer chasing cheap labor worldwide over investing in automation at home, the countries that lead in the robot race chose to try to keep their manufacturing industries, using labor-saving technology to offset the cost of high wages instead of offshoring manufacturing to low-wage foreign countries. Deindustrialization is a political choice by national governments and corporations, not the inevitable result of abstract forces such as globalization and technology.
The United States’ partial deindustrialization, made possible by corporate access to low-wage foreign labor pools, transformed American urban economies in separate ways. Some deindustrialized parts of the Midwest, Northeast, and South have become economic and social disaster areas, with rusting factories and high rates of unemployment and social pathologies such as opioid addiction.
The number of “tech” jobs has increased slightly, but many of these good knowledge jobs have been captured by H-1B holders and other guest workers and green card holders. Facebook, one of many offenders in the tech community, recently paid the US Justice Department $14 million to settle a lawsuit claiming that it favored foreign workers over American workers.14 And good tech jobs in the US may be lost if US-based multinationals outsource innovation and manufacturing and services to workers in China, India, and other countries.
The Limits of Urban Triumphalism
Ultimately, Silicon Valley symbolizes the US transformation from a dynamic manufacturing country into a postindustrial economy dominated by rentier interests. Originally, many computer components were made by workers in the Bay Area.15 Offshoring production in search of cheap labor has turned hub cities such as San Francisco and New York into something like high-end resort communities. Tech entrepreneurs and tech firms licensed their innovations, and money flowed to them from around the world and flowed out via their spending on luxury consumer amenities, maintained by an increasingly hard-pressed urban service workforce.
Society in the Bay Area and similar metro areas has polarized between tech billionaires and well-paid professionals and a low-wage service class that includes mass unskilled immigrants, many of them unauthorized: nannies, house cleaners, gardeners, food-truck operators, and Uber drivers. Half the nation’s homeless population lives in the Golden State, many concentrated in disease- and crime-ridden tent cities in either San Francisco or California’s largest city, Los Angeles. The City by the Bay, notes the Brookings Institution, has the second-highest level of inequality in the country.16
Until recently, this emerging urban pattern of the 1990s and 2000s was viewed as the pinnacle of social and economic innovation, the model of future society in the US and other developed countries. Expanded educational opportunities would help disadvantaged Americans join the privileged tech and global services elite. Those who remained in the working class, now dominated by menial service workers, would be helped not by higher wages but by more generous government transfer payments. The “winners” in the “new economy” would be taxed to subsidize the “losers” via a slightly expanded welfare state.
Something like this remains the consensus in the Democratic Party, to judge by the Biden administration’s ill-fated Build Back Better spending plan in 2021. Following the separate passage of a bipartisan ordinary infrastructure bill, the second bill in the series consisted of investment in renewable energy and “social infrastructure.” The latter was defined by programs such as promotion of urban densification and federally subsidized commercial and public day care. The presumed beneficiaries of urban densification would be urban real estate interests and well-educated professionals, particularly young, single, and often childless professionals. The service class would perform many functions that once took place in the household—cooking meals, walking dogs, cleaning residences, and taking care of offspring.
This future vision saw big cities dominated by an urban, college- educated “creative class”; below this, barely noticed, would be service workers who often could barely afford to live in urban areas. Many of these service workers joined the older cohorts of professionals in the suburbs, although often in vastly distinct locations. For a long time, except in a few cities, suburban growth has outpaced downtown growth, and exurban growth has been fastest of all, with non-white homeowners growing as a share—partly because of the rising cost of living in urban cores.17
The Pandemic’s Impact
The COVID-19 pandemic and the accompanying economic and social disruptions, including the largest urban homicide and theft waves in generations, may have accelerated the decline of the urban cores that only recently were seen to represent the future. During the pandemic, many city dwellers relocated from downtowns to suburbs or the countryside, and some may never return. In the next decade or two, the reputations of big cities such as New York and San Francisco may be more like they were in the grim 1970s and 1980s than the booming 1990s and 2000s, identified with scenes of urban squalor: boarded-up stores, empty buildings, homelessness, and crime.
Necessity is the mother of invention, and the lockdowns caused by the pandemic accelerated the adoption of technologies that enable telecommuting, such as Zoom. If a chronic threat of global pandemics makes air travel and large public gatherings permanently more onerous, at least some of the replacement of face-to-face contact by remote interaction may become permanent.
If this occurs, then there will be major implications for urban workforces.18 There would be continued high demand for workers deemed “essential” during the COVID-19 pandemic, such as those in factories and warehouses, logistics and transportation, nursing, health care, and elder care. But if downtown business districts never fully recover, many of the low-end service-sector jobs that depended on them may disappear permanently. The costs of such disruption will fall heavily on women, non-white workers, and immigrants, who are overrepresented in nonessential but “frontline” services, compared to essential workers, a demographic that resembles the US workforce.19
Some of the displaced urban service workers might find new jobs in a permanently expanded home-delivery sector like the one that sprang up during the epidemic. But many of these jobs, such as delivering groceries to professionals who telecommute, are low-skilled “Mechanical Turk” jobs that could be replaced in time by delivery robots with sufficiently advanced artificial intelligence.
As more consumers prefer direct delivery to their homes or post office boxes of goods that are ordered online, the decline of shopping centers and malls, already underway before the pandemic, could accelerate. And even where business or pleasure involves a return to older patterns, automation may eliminate formerly abundant low-wage urban jobs— with kiosks and delivery services replacing waiters and servers in many restaurants, for example.
Despite such incidental disruptions, service-sector automation should be welcomed, not feared. It can boost the US economy’s overall productivity. Indeed, the abundance of low-wage labor in the US in the past generation— resulting from numerous factors, including the replacement of better-paid jobs by worst-paid ones, the decline of unions, the large-scale entry of women into the workforce, and high levels of unskilled immigration— may have slowed productivity-enhancing automation in the US by reducing the incentive for employers pressured by high wages to invest in labor- saving technology.
Indeed, in some low-wage sectors, an abundant supply of cheap labor may have generated the demand for it in the past generation. If demand is elastic, it expands as the price goes down and shrinks as the price goes up. Elastic demand is associated with luxuries, which many people willingly forgo if the price increases—for example, cooking at home if restaurant prices rise.
Many of the bad jobs created in the past generation in the US have been in elastic-demand luxury sectors that pay low wages—think of neck massages in airports or shopping malls. Given the large population of low-wage, immigrant workers, what were once labor-intensive amenities for the rich became affordable for many middle- and working-class people in the past generation. An example is the day spa, a spin-off of the traditional resort spa, often catering to professionals and workers who work nearby in downtown areas. One industry survey found that 46 percent of day spa clients made less than $35,000 annually, while 65 percent made less than $65,000.20
The Future of the Service Class—and That of Our Cities
A huge, low-wage, urban service sector, then, was not inevitable but developed because of the confluence of several factors. These include the concentration of the free-spending affluent and rich in a small number of cities, much of it “bubble” wealth flowing in boom times to rentiers from assets with inflated values such as stocks and real estate; stagnant or falling wages for male breadwinners, which forced many women reluctantly into the low-wage workforce; the lowering of wages because of de-unionization; and the replacement of full-time workers by more poorly paid contractors or contingent workers in an insecure, low-wage “precariat” class. Essential to the process has been the large numbers of unskilled immigrants desperate enough to work for low wages and clustered in the same cities as affluent professionals, managers, and rentiers.
Had any of these macroeconomic or microeconomic variables been different—tighter labor markets because of lower immigration or more unionization, for example, or a Federal Reserve policy that preemptively priced asset bubbles instead of accommodating them—the pattern of service-sector spending and employment alike in American metro areas might have been quite different in the past generation.
What if the “bubble era” urban economy of the 1990s and 2000s never returns? With less demand for their services from downtown professionals, managers, and rentiers, many members of the “frontline” service-sector working class, both nonessential and essential, may have to find new jobs.
Some former downtown service workers may move to cheaper suburbs or exurbs in the same state or other states to follow affluent customers, enjoy a lower cost of living, or both. Food trucks and downtown day spas and trendy restaurants may shrink, shedding labor to sectors less dependent on concentrations of affluent downtown consumers. In a more decentralized America, more workers may be employed in delivery and transportation services that disproportionately serve telecommuting professionals.
Some of the current mass withdrawal from the labor market may be voluntary, particularly among families. Polls show that for most Americans, the one-earner family, in which one parent stays home to care for young children or works only part-time, is the ideal. According to one 2021 survey, the preferred option for childcare among Americans is one full-time stay-at-home parent (40 percent), followed by two part-time working parents (18 percent). Institutional day care is preferred by only 12 percent. Regarding day care, there is a class divide, with 22 percent of college-educated parents preferring full-time, center-based day care, compared to only 10 percent of parents without college degrees. But even college-educated parents prefer “flexible work + shared child care” (28 percent) or one full-time at-home parental caregiver (24 percent) to institutional day care (22 percent).21 If technological progress continues to make many necessities cheaper, wages in general increase, or both occur, then households can choose to maintain the same standard of living with only one full-time worker rather than two.
For self-interested reasons, employer lobbies that prefer a seller’s market in jobs and a buyer’s market in labor treat any reduction in the share of adults employed as a disaster. But more single-earner couples and fewer two-earner couples might shift consumer demand from some areas to others, without necessarily lowering overall consumer demand. Similarly, more household do-it-yourself production rather than mass sales, enabled by ever-cheaper technologies such as rapid prototyping, also known as 3D printing, would increase demand for appliances and inputs, even as it reduced demand for finished goods and services.
For those who remain in the workforce, out of choice or necessity, there is no reason to fear a lack of jobs. Since the 1960s, there have been periodic alarms about the prospect of mass unemployment resulting from automation. These fears have never materialized and are unlikely to do so in the future.
To be sure, automation, such as mechanization, can disrupt employment in particular industries. Desktop computers eliminated jobs for many typists and secretaries, just as cars and trucks eliminated equine- drawn carriages and wagons. But to date there have been no examples of mass unemployment caused by machines replacing human beings. Beginning in the 19th century, all major depressions and recessions have been because of financial panics or stock market or real estate bubble collapses with economy-wide contagious effects, not because of labor- saving technology.
One reason for the absence of mass technological unemployment is that the same substitution of machinery for labor that displaces some workers lowers prices for goods or services. This allows consumers to consume more of the now-cheaper goods or services if they choose (“elastic demand” again). Alternatively, consumers can use the savings from cheaper material goods to spend more on quality-of-life services. As incomes increase be- yond a point, the consumption of material goods levels off, and the affluent and rich spend a greater share of their income on luxury services provided by other people. If automated grocery stores and store delivery reduce grocery prices, then consumers might spend the savings on labor-intensive amenities, such as catering, personal gyms, physical therapy trainers, personal shoppers, and other services formerly affordable to only the rich.
Big Opportunity Ahead: Health Care
The ultimate quality-of-life service is health care. The late economist Robert W. Fogel concluded:
The income elasticity for health services is calculated at 1.6, meaning that income expenditures on health care in the U.S. are likely to rise from a current level of about 15 percent to about 29 percent of GDP [gross domestic product] in 2040.22
While the numbers are too specific, the point is confirmed by observation. The wealthy voluntarily spend a much greater percentage of their incomes on health care than do the poor and with good reason; health is the good that makes the enjoyment of all other goods possible. It makes sense that societies, like individuals, will spend more on health care out of choice as they become more affluent, particularly if the cost of other goods such as food, clothing, housing, and transportation continues to go down thanks to technological productivity growth.
This suggests that the influential futurists of the 1990s chose the wrong “industry of the future.” The real industry of the future is not “tech,” defined as software innovation, but medicine.
The medical industry could be as central to 21st-century industrial nations as the automobile-manufacturing complex was to the 20th century. From the point of view of demand, it is even better. Most households do not need more than a few cars, but the demand for medicine will keep going up as the price falls.
Moreover, the medical industry combines many diverse industries. Most medicine belongs to the domestic-traded service sector, employing workers at all skill levels, including nursing aides, nurses, nurse practitioners, internists, and specialists. But some parts of the medical industry are traded-sector occupations with potential global markets—for example, medical tourism, medical manufacturing for prosthetics, and even perhaps global and national telemedicine and long-distance surgery. Best of all, the medical sector’s profitability ensures that, unlike in other sectors, even unskilled workers can enjoy high wages, because of government regulation or pressure from organized labor.
The functional equivalent of 20th-century Detroit may be something like the Texas Medical Center in Houston. This suggests that industrial policy—now being rehabilitated, after the neoliberal consensus stigmatized it for a generation—has a place in the service sector and traded-sector industries such as manufacturing. The goal of a service-sector industrial policy should be to increase wages and productivity in low-wage sectors such as retail and hospitality, even if this causes those sectors, as they became more efficient, to shed labor. The displaced labor can be absorbed in other industries such as health care, which benefit from rising demand, rising productivity, and a mix of jobs at every skill level.
Beyond the “Knowledge Economy”
In hindsight, Silicon Valley turned out to be an outlier rather than the future of the American economy. The myth of the “knowledge economy” was a false dawn. The actual future of American employment—and the American city—will be shaped by technological and global trends but could be far brighter given the right domestic program, including better industrial and labor policies. Advanced technologies provide the tools nations can use to structure their own economies, but different countries can use the same tools differently. The structure of the world economy constrains, but does not determine, the different options that nation-states can choose.
The future of the urban workforce does not need to be imposed on Americans by technological or economic forces beyond their control. For better or worse, it will be made in America by Americans, and it should be aimed at restoring the prospects for those whose labor has sustained us before and particularly during the pandemic and will continue to do so in any conceivable urban future.