The Next Industrial Revolution

An interview with the Economist columnist Ryan Avent on his new book about how technology will change the labor force

Brian Snyder / Reuters

A “crisis of abundance” initially seems like a paradox. After all, abundance is the ultimate goal of technology and economics. But consider the early history of the electric washing machine. In the 1920s, factories churned them out in droves. (With the average output of manufacturing workers rising by a third between 1923 and 1929, making more washing machines was relatively cheap.) But as the decade ended, factories saw they were making many more than American households demanded. Companies cut back their output and laid off workers even before the stock market crashed in 1929. Indeed, some economists have said that the oversupply of consumer goods like washing machines may have been one of the causes of the Great Depression.

What initially looked like abundance was really something more harmful: overproduction. In economics, as in anything, too much of a good thing can be problematic.

That sentiment is one of the central theses of The Wealth of Humans, a new book by the Economist columnist Ryan Avent about how technology is changing the nature of work. In the next few years, self-driving cars, health-care robots, machine learning, and other technology will complement many workers in the office. Counting both humans and machines, the world’s labor force will be able to do more work than ever before. But this abundance of workers—both those made of cells and those made of bits—could create a glut of labor. The machines may render many humans as redundant as so many vintage washing machines.

Once again, what once seems like abundance will instead be over-supply: The machines may invent their makers out of work.

Last week, I spoke with Avent about his book, how his theories might help to explain the 2016 election, and the future of working. The following conversation has been edited for clarity and concision.

Derek Thompson: In classic Economist style, your title, The Wealth of Humans, is doing double or triple duty. First, it’s a play on Adam Smith’s The Wealth of Nations, and indeed there’s a lot of Smith in here. Second, it’s a book about the most common definition of wealth, money, and how it might be earned and distributed in the future. Third, it’s about Merriam-Webster’s second definition of wealth, which is a surfeit, a surplus, and your argument is that we may be entering a world with too many workers. Anything I’m missing?

Ryan Avent: Those were the ones I had in mind. There may be others lurking.

Thompson: There is an ongoing debate about whether technological growth is accelerating, as economists like Erik Brynjolfsson and Andrew McAfee (the authors of The Second Machine Age) insist, or slowing down, as the national productivity numbers indicate. Where do you come down?

Avent: I come down squarely in the Brynjolfsson and McAfee camp and strongly disagree with economists like Robert Gordon, who have said that growth is basically over. I think the digital revolution is probably going to be as important and transformative as the industrial revolution. The main reason is machine intelligence, a general-purpose technology that can be used anywhere, from driving cars to customer service, and it’s getting better very, very quickly. There’s no reason to think that improvement will slow down, whether or not Moore’s Law continues.

I think this transformative revolution will create an abundance of labor. It will create enormous growth in [the supply of workers and machines], automating a lot of industries and boosting productivity. When you have this glut of workers, it plays havoc with existing institutions.

I think we are headed for a really important era in economic history. The Industrial Revolution is a pretty good guide of what that will look like. There will have to be a societal negotiation for how to share the gains from growth. That process will be long and drawn out. It will involve intense ideological conflict, and history suggests that a lot will go wrong.

Thompson: Even I would admit that is a weird time to predict the end of work, considering that the unemployment rate has been at or under 5 percent all year, the private sector in the U.S. has created jobs for record-high 77 consecutive months, and wages are actually rising at their fastest rate since the Great Recession.

So what is the best evidence that your prediction is plausible?

Avent: I would say the best evidence comes from the wage growth numbers. I know we’ve experienced an uptick in recent months, but we’re seven years into the recovery and still well short of the level of nominal wage growth we would expect, even compared to recent disappointing recoveries. In the bigger picture, for a lot of middle-skilled workers, especially men, you have stagnating wages for several decades. Apart from the top 1 percent, a lot of people are having a lousy time.

If you look at the experience of rich countries across the world, you see there is a tradeoff between wage growth, productivity, and employment growth. Employment in Britain is at an all-time high, and wage growth there has underperformed America and most of Europe. This suggests that the main way that employers are using people in countries like the U.K. is to use them to do low-productivity work.

Thompson: There is a familiar story of technology and the labor force that one might call the “we used to” story. We used to work on farms, we used to work in textiles, we used to work in factories … What’s the next chapter of the “we used to” story? What sector currently employing a lot of Americans is the lowest-hanging fruit for disruption?

Avent: Driving is certainly an area where we’ve seen more rapid progress than I would have guessed. Truck drivers, bus drivers, and train drivers have pretty good pay and those account for millions of jobs. Most importantly, there seems to be an interest among companies employing those workers to bring [the tech that would replace humans] forward. In the long run, I’m optimistic for technology to transform health care, but that’s a harder sector to disrupt.

Machine intelligence will be applied in ways we cannot imagine yet. One example is talking. Today, if you have a problem with a car company, you might end up conversing with a bot over the phone. Those are conversations that we thought weren’t automatable that are now. We used to employ a lot of people to talk to people and people have those conversations with bots.

Thompson: At the moment, there is some evidence that wages are rising fastest at the bottom, which is an interesting challenge to these theories. You’ve predicted that wages won’t rise for a substantial share of middle and low-skill workers, because there is an abundance of labor. But you also predict that high wages for easily automated jobs will be a big fat target for automation, because employers will want labor-saving technology wherever they can save the most money.

So, I wonder, do you regard rising wages for fast-food workers today as a challenge to your theory, or a development that will hasten the automation of fast-food joints, because when McDonald’s workers are earning high wages, it’s more tantalizing for management to replace them with machines?

Avent: It will be interesting to see how increases in the minimum wage affect this. You see stories of robotic burger-flippers and people ordering food through iPads. It’s possible you see more of that as minimum wages rise. It seems like there is technology waiting on the shelf to displace that work.

I think I would say: If in 10 years time, workers in those jobs have received a substantial raise relative to current levels, and employment has not fallen, and productivity has not gone up, that would be evidence that I’m wrong. But my expectation is that as these workers become more expensive, you’ll see more interest in the technology that could displace them.

I should add that that would be a very good thing! We want businesses to invest in tech that makes the economy more productive.

Thompson: Your book is very good on the intersection of two ideas that don’t often interact in economic analysis, which is the future of work and housing policy. It is ironic that liberals, who in the abstract support more inclusive immigration and shared wealth, often live in coastal metros areas that are exclusive by design—they are built around water, limit housing height, and declare certain zones out-of-bounds for further construction. As you point out in the book, one of the tallest buildings in New York City is a residential tower on Park Avenue that is home to a stack of billionaires who, although they could live in any ZIP code on the planet, have chosen to live on top of each other, like candies in a Pez dispenser.

I can imagine a future where these rich liberal cities might also be on the frontline of the coming automation wave. It seems to me that technology adoption for services like Uber spreads fastest among densely populated areas with young, educated workers who are early adopters. That’s cities. Plus, labor-saving technology is most tantalizing where wages for low-skill work is highest. That’s cities, too. Might these rich cities be canaries in the coal mine for technology displacing human workers?

Avent: That’s an interesting question: I think I would say yes. Big cities like London, New York, and San Francisco play out in miniature the debates we’ll see at a national level. They’re generating phenomenal wealth, and people who don’t have access to it can be left behind.

These cities tend to be early adopters of Uber and other sharing economy apps, which are systematizing jobs in ways that make them more automatable. For example, once you take the brain out of the driving, it’s just a person following a map, and it’s easy to imagine a machine just following a map. I could also imagine that in these cities, in-person contact becomes something like a luxury good.

Thompson: Like the phenomenon of $120,000 nannies and Latin tutors on the Upper West Side.

Avent: Yes. The very rich will still want people, their own personal shoppers and assistants. Being able to retain human labor would be a sign that you’re wealthy. So even in a future city that had a lot of laborers replaced with technology, you might still have artisanal service sector workers.

Thompson: One of the popular solutions to this problem is the idea of a universal basic income, or UBI—a bare minimum that every adult would be owed that is paid out of this new wealth. What do you think of UBI?

Avent: Right now, our safety net is mostly insurance for the young and the old and people with bad luck. The expectation is that most people pay into it and therefore they expect insurance when they get old or sick. A universal basic income is a totally different social contract. It says that, on a permanent basis, a large class of people will probably be subsidized by a different class. That’s a much trickier thing politically, and it raises questions about the value that they are contributing to society.

Thompson: This question of who gets what in America and who is worthy of a piece of our national wealth seems in many ways to be at the heart of Bernie Sanders and Donald Trump phenomena, don’t you think?

Avent: Bernie Sanders and Donald Trump are two sides of the beginning of this social evolution. Bernie is pushing along the direction of, "let’s distribute more." Trump is pushing along a related direction which is, "let’s exclude others who are not like us." The book talks at the end about how redistribution creates political pressure to exclude those who don’t belong, who aren’t like the majority. The rise of nationalistic political figures in America and in Europe is related to the idea that the majority wants to draw the circle of society closer.

Thompson: I think that too often work is defined in these conversations as a narrow exchange: employee labor for employer money. But as you write—not only in this book, but also in your essay on hard work—a job is so much more than that. It’s a social network, it’s a distraction, it’s a way to fill the day, it’s a source of status and pride and ownership. Even if the number of salaried jobs as a share of the labor force declines, this is something we really don’t want to lose.

Avent: People enjoy work. Even those who don’t enjoy what they do enjoy the feeling of agency and being able to provide for others. For a world to work where a universal basic income accounts for the bulk of the consumer spending for many people, something else needs to account for the social side of work. It is disappointing to think that we’d have to create make-work for people, but it may be the hard truth.

We’re a long way away from that world. What comes next would be higher wage subsidies and in-kind benefits, like tuition-free college or subsidized health care. But it’s coming, and the debate will be: If we’re going to pay people to do work that isn’t necessary, who do we let into the system? Who is allowed to benefit?