Evil Geniuses: The Unmaking of America: A Recent History (40 page)

In the same chapter, near the end of
Democracy in America,
Tocqueville worried that “the manufacturing aristocracy which is growing up under our eyes is one of the harshest which ever existed in the world,” and that America’s new capitalist apparatus “impoverishes and debases the men who serve it, and then abandons them.” He issued a warning:

The friends of democracy should keep their eyes anxiously fixed in this direction; for if ever a permanent inequality of conditions and aristocracy again penetrate into the world, it may be predicted that this is the channel by which they will enter.

*1
In fact, all the Nordic governments are certified non-socialist-oppressors. According to the annual ranking of countries by the conservative Property Rights Alliance, a spinoff of Grover Norquist’s Americans for Tax Reform, Finland ranks first, Sweden seventh, Norway eighth, the United States twelfth, and Denmark thirteenth. And on the Heritage Foundation’s Index of Economic Freedom, all the Nordics and the United States are clustered between eleventh and twentieth.

*2
By the way, even controlling for other exceptional U.S. facts like obesity and guns, people in other rich countries are living longer than Americans, and life expectancy for nonwhite Americans has been increasing
faster
than for whites.

This book is subtitled
A Recent History.
So far I’ve focused mainly on the last fifty years, and on how we let a powerful clique and its enablers turn a quintessentially modern and reasonably fair political economy that led the world into a freakishly old-fashioned and unfair one.

But what now? What happens
next
to the rules of how Americans work and earn a living and help (or ignore) one another, during the next few years and over the next few decades? How should the United States be fixed? Will it be fixed? Can it be fixed?

This final section looks forward, anxiously and hopefully. The continuing transformation of work (and life) by artificial intelligence and automation is an absolute given, for better or worse—potentially much better or much worse, depending on whether we choose to use technology for the benefit of most of us or only a few of us. Technological progress and social progress are two different things. As always, how or if the former leads to the latter will be for politics to determine.

When I started this book, I worried that my strong sense that we’d arrived at a historic crossroads equivalent to the 1770s or 1860s or 1930s, America’s Fourth Testing, might seem like a stretch. By the end, I was no longer concerned about overstatement, particularly after the pandemic arrived—a new virus requiring new behaviors and policies, work and life suddenly more than ever dependent on the Internet, government failure by leaders ideologically dedicated to undermining government, maximizing corporate profit at all costs, and an American hyperindividualism whipped up by the right that makes a huge common problem harder to solve.

So here we are, abruptly shoved into a new age, no longer merely on the threshold of a strange future but altogether
in
it. The year 2020 will become one of those dates ripe with significance, like 1945 or 1914. We don’t yet know exactly what era is ending and what is beginning. I think we could prevent America from turning into a permanent dystopian horror show. We might even manage to make it better than it was before the evil geniuses destroyed so much. Either way, we’re never returning to normal, life just the way it ever was. And the future is going to be outlandish.

Scholars and journalists and politicians debate whether, after 1980, automation or overseas manufacturing was more responsible for eliminating U.S. jobs and keeping wages lower. It’s certainly easier psychologically and politically to
blame
foreign workers (and immigrants) and foreign trade deals than to blame computers and other machines. But the question is now mostly academic.

Starting in the 1990s, technology and globalization worked as a kind of job-killing tag team. When cheaper foreign goods came flooding into the United States, one way U.S. manufacturers lowered their costs to compete was by automating, but the Internet also made foreign trade and overseas manufacturing much easier. In any case, the large-scale replacement of U.S. workers and jobs with cheaper workers abroad essentially finished a decade ago, when the fraction of Americans working in factories dropped below 10 percent and for practical purposes there were hardly any more jobs to offshore. Likewise, starting in the 1990s, white-collar work was moved overseas (such as customer service jobs in call centers) but also increasingly rendered obsolete by digital technology, as the careers for Americans in music and newspapers and travel booking and tax preparation and video rental and other retail businesses came to dead ends. Economic efficiency wins, by whatever means. In the first decade of the new century, another 5 million U.S. manufacturing jobs went
pfffft
even though more stuff was actually being manufactured in the United States. Unlike the way things were set up before the 1980s, all that increasing efficiency and productivity benefited investors and high-end professionals, not most workers.

In the early 1980s, when the outsourcing and offshoring of American jobs had barely started and those terms were just being coined, one esteemed economist suggested that this economic dislocation would be a temporary phase. But he didn’t mean in a good way. Back then, Chinese factory workers were paid the equivalent of two dollars a day. As China and other developing countries developed and started paying higher wages, he predicted, the automation of manufacturing would accelerate, meaning that work would “tend to move back to the advanced industrialized countries.” And so it is several decades later, now that the average Chinese manufacturing worker makes $11,000 a year, approaching the U.S. minimum wage. For now, there are still cheaper workers in poorer countries. But as the cost of overseas labor keeps catching up with (stagnant) U.S. wages, the logic of efficiency dictates the next phase: just as American workers were replaced by foreign workers who could be paid less and treated worse, now foreign workers will be replaced by unfeeling, uncomplaining machines.

Wassily Leontief was stunningly prescient in that 1983 paper about the future of technology and work, which he wrote ten years after he won the Nobel Prize for economics.
^*1
Back in 1949, the middle-aged Leontief had started using one of the earliest computers for his research at Harvard. While early “computerization and automation” had helped increase wages and “labor’s total share in the national income,” he wrote in 1983, he predicted that before long those same forces “are likely to begin to operate in the opposite direction”—which was exactly what proceeded to happen.

Back in the early 1980s, the share of the U.S. economy devoted to providing services rather than making (or mining or growing) physical things was still just a bit more than half, but “after 60, 70, or even 90 percent of total output is provided by service industries”—it’s now hitting 80 percent—Leontief predicted that “continued computerization and automation will ultimately prevail” and push down wages for service workers as well as manufacturing workers. And so it has. What’s more, because “any worker who now performs his task by following specific instructions can, in principle, be replaced by a machine,” that was bound to happen and accelerate, so “labor will become less and less important. I do not see that the new industries can employ everybody who wants a job.”

For a few years, Leontief had refrained from publicly sharing his dark analogy about the inevitable process he foresaw, but with the mass computer age arriving—10 percent of Americans owned a PC and 1 percent were on the Internet in 1983—he decided on candor. What happened to horses in America during his lifetime, he wrote, when we went from owning 26 million in 1915 to owning 3 million in 1960, would happen to human workers as we entered the twenty-first century. “A reduction of oats rations allocated to horses could delay their replacement by tractors,” he said of horses in the old days. “But in the case of human rations, or wages,” that would work “only in the short run” because

in the long run, technology will always be introduced, whether quickly or slowly….Had horses had an opportunity to vote and join the Republican or Democratic Party, that solution would’ve been very different from what it actually was. Had we for instance wished to maintain the 20 million unemployed horses by putting them to pasture we certainly could’ve done so.

The proposition that the worker who loses his job in one industry will necessarily be able to find employment, possibly after appropriate retraining, in some other industry, is as invalid as would be the assertion that horses who lost their jobs in transportation and agriculture could necessarily have been put to another economically productive use.

In the thirty-seven years since Leontief wrote that, of course, America has indeed put increasingly uneconomic workers on short rations. And although there’s still a lot of optimistic, hand-waving conventional wisdom about the future of jobs in the new AI era, a national conversation about putting workers out to pasture with universal basic incomes has begun.

Automation and computers replace human workers in all kinds of ways, some more obvious and visible than others, but the process of actual robots taking over the jobs of Americans is still in its early days. Maybe a million U.S. workers—machinists and welders and the like—have already been replaced by robots. But beyond those direct replacements of skilled workers, two MIT economists recently found that during the first wave of industrial robots in the 1990s and 2000s, each robot installed led to the loss of a half-dozen jobs. The cost of robots is dropping, and the number installed in American factories has been doubling every few years and has passed a quarter-million. But our “robot density” is still less than a third of South Korea’s and is also much less than that of Japan and the advanced European countries. At the end of 2019, there were still millions of U.S. manufacturing jobs waiting to be automated out of existence by robots and other machines.

The easy summary of what’s afflicted our political economy the last forty years is economic inequality and insecurity, fortunate people at and near the top getting paid more and more and remaining highly employable, but no such luck for almost everyone else. Underlying the growing differences in income and wealth and security are more complicated changes in
how
Americans are able to earn livings.

The central problem is that since the 1980s, jobs with good salaries and benefits in the great big
middle,
both blue-collar and white-collar—machine operators, mechanics, clerks, assistants, bookkeepers, salespeople—have disappeared ever faster. And 90 percent of those jobs, according to a new study, disappeared right after recessions, each recession the prod or pretext for businesses not only to get rid of some current employees but also to eliminate those positions permanently. (Watch for that to happen again in 2021 and 2022.) That sloughing of routine jobs has been accelerated by computers. The impact of automation so far hasn’t much reduced the overall
number
of American jobs, but it has forced redundant millions into
worse
jobs—jobs that require fewer skills and pay less, like a person I know who went from being a skilled printer to working as a shopping cart wrangler. According to a 2018 study by the MIT technology and labor economist David Autor, “over the last 40 years, jobs have fallen in every single industry that introduced technologies to enhance productivity,” and “automation pushes workers to the less productive parts of the economy.” That is, they’re pushed out of well-paid factory and office occupations into work that can’t (yet) easily be automated, jobs such as security guards and hospital orderlies, waiters at Outback and cart guys at Home Depot.

Meanwhile, however, at each
end
of the economic ladder, the six-figure marketing executives and engineers and seven-figure bankers
and
their full-time and part-time servants and attendants, still have work aplenty and get increases in pay. This is what people mean when they talk about the job market having been “hollowed out.”

In particular the
automators,
the college-educated workforce employed by digital companies that didn’t exist until recently, are doing fabulously. At Facebook, Google, Netflix, and thirty or thirty-five other companies among the largest five hundred corporations, the
median
pay is now over $150,000. The problem for everyone else who needs a job is that most digital companies are so phenomenally efficient they don’t require many workers.

I was amazed when I compared the revenues and workforces of two of the biggest companies with those of two of their goliath predecessors back in the day, General Motors and AT&T. In 1962, when GM made most of the cars sold in America and AT&T manufactured and operated all the telephones, the two together employed more than a million people, one out of sixty American workers, each of whom generated revenues for the companies equivalent to $170,000. Today Apple and Google have revenues
and
profit margins twice as high as GM and AT&T had in 1962, but together employ only a quarter-million people, just one out of six hundred American workers. Thus those newer companies, in constant dollars, collect twelve times as much revenue per employee, and they earn twenty times as much profit in all. The market value of Apple, Google, Facebook, and Microsoft combined is nearly 10 percent of the value of all public companies, yet those four employ only a small fraction of one percent of the U.S. workforce.

Some commentators and scholars have ventured a glass-half-f take on these new small but well-paid digital workforces—that the affluent employees effectively create lots of jobs by living so lifestylishly. According to one estimate, each job at a tech company results in five additional jobs elsewhere—two financial advisers or architects or therapists and the like, plus three nannies and housekeepers and gardeners and dog walkers. There are now more than 4 million of these so-called “wealth workers” in the United States, which doesn’t even include the additional millions of drivers or waiters and bartenders who, except during pandemics, also service the well-to-do.

It’s another way our political economy and society have been dragged back in time. Not only have we let economic inequality and insecurity revert to the levels of a century ago, we’ve also shrunk the middle class down to the size it was in the old days
and
reconstituted a vestigial caste to make life even easier for the well-to-do. A century after it began disappearing in America, the servant class is back.

Speaking of Google, as a result of Europe’s successful adoption of good old-fashioned mid-twentieth-century-U.S.-style scrutiny of big companies, the antitrust enforcers of the E.U. in three separate actions from 2017 to 2019 fined Google $9 billion for anticompetitive practices. If Google finally loses its appeals and pays all those fines, it would be a serious hit for the company, about a tenth of its annual profits for those three years.

In America, meanwhile, we’ve only just begun imagining how we might make Google and the other new digital mammoths serve the U.S. public interest. Google and Facebook are exactly the kinds of monopolistic enterprises for which America began enacting antitrust laws at the end of the nineteenth century.

But as I’ve said, over the last half-century, antitrust violations were redefined to mean
only
corporate actions specifically intended to raise prices for consumers. Most of Google’s billions of users pay nothing directly to Google for its services (although like me, millions have surrendered to the monopoly and recently started paying twenty-five dollars a year for extra email storage). So Google’s economic power is a new, befuddling kind—
everyone
is a user, but their real paying customers are all other businesses. According to the legal scholar Eric Posner and the economist Glen Weyl, both free-market superenthusiasts, “antitrust authorities are accustomed to worrying about competition [only] within existing, well-defined, and easily measurable markets”—that is, not within the new market of digital advertising, which now constitutes
most
advertising. Back in the pre-Internet, pre-cable day, the closest equivalents were the three (highly regulated) TV networks, which together took in maybe 12 percent of all U.S. ad spending. Google and Facebook get almost
two-thirds
of all digital ad revenue, which is why those two companies alone are worth $1.5 trillion.