Cities Used to Help All Americans Find Prosperity. Not Anymore.

People picnicking in an urban park.
The place for you? Maybe not.
Robert Bye/Unsplash

One way economists have made sense of the growing divide in the United States is by looking at mobility, or the lack thereof: 11 percent of Americans moved last year, the lowest rate on record.

At the same time, economic conditions between cities, and between cities and rural areas, have diverged over the past three decades. The Brookings economists Clara Hendrickson, Mark Muro, and William Galston recently summarized the past decade in employment growth, which has occurred almost entirely in big cities.

A graph showing divergent employment outcomes for cities and rural areas since the 2008 recession.
Smaller communities still haven’t recovered from the Great Recession.
Mark Muro/Brookings

Data like these have given rise to what we might call the Greyhound theory of economic growth: Instead of trying to help struggling places, give the Americans who live there a chance to get to a thriving city. Even Donald Trump outlined a version of this in the summer of 2017. “When you have an area that just isn’t working like upper New York state, where people are getting very badly hurt, and then you’ll have another area 500 miles away where you can’t get people, I’m going to explain, ‘You can leave,’ ” the president told the Wall Street Journal.

One controversial aspect of that advice is that it leaves the struggling places for dead, though others argue population decline need not be tied to economic decline. In the keynote lecture at the American Economic Association conference in Atlanta last week, MIT economist David Autor raised another issue: Moving to the big city might not be a golden ticket after all.

Looking at population density, wages, and worker education, Autor graphed a striking half-century transition in what constitutes work in American cities. In 1970, a 30-year-old college-educated worker and 30-year-old high-school dropout both earned much more money in more densely populated counties. Since then, the paths of the “college or greater” group and the “high school or less” group have diverged. Highly educated workers have an even more pronounced “urban wage premium” than in 1970. For workers with a high school education or less, the urban wage premium has virtually vanished. Moving to the city, in short, pays off handsomely if you have a college degree—and hardly at all if you didn’t graduate from high school.

Graphs showing divergent outcomes over time in cities for educated and less educated workers.
Some college or greater in blue, high school or less in red. Wages have traditionally been higher in more densely populated counties. Now, that’s only true for more educated workers.
David Autor/MIT

“Why didn’t these laid-off manufacturing workers buy a bus ticket to Manhattan? This explains that,” Autor told me on Wednesday.

Those figures may underestimate just how limited the big city’s opportunities are for low-skill workers. In some ways, a city isn’t a bad place to be poor: Social services are often more easily accessible than in rural areas, and car ownership is unnecessary. But that effect is likely offset by the enormous burden of big-city housing costs, which further decrease the promise of the Greyhound theory.

But Autor’s point is more comprehensive than that. Building more housing isn’t rocket science, but resurrecting the mid-skill work that made American cities engines of social mobility is more complicated, and maybe impossible. Driven by decentralization, deindustrialization, union-busting, outsourcing, and computerization, a whole host of urban middle-class jobs (manufacturing, clerical work, administrative jobs) have vanished from American cities. Workers have generally moved up the skill ladder into professional, technical, and managerial roles—a trend that matches Americans’ increasing educational attainment.

But not all workers. Young men make less money, at the median and adjusted for inflation, than they did in 1975. Less-educated workers, and particularly men, are moving down the employment bracket, out of mid-skill work and into service jobs like hospitality, security, food service, and manual labor. When these workers arrive in cities, Autor says, they’re actually finding a smaller share of mid-skill work than they would in small towns and rural areas. They wind up working low-skill jobs with low wages and little opportunity for advancement.

This research does not fly in the face of the great 20th-century assumption that the metropolis is a place where fortunes are made: If you’ve got an education, you’re still going to make big bucks in the big city. But it does illustrate how limited that transformative capacity has become at the bottom of the ladder.

One interesting thing about this research (you can review the slides or watch the lecture here) is that it doesn’t account for regional inequality. In the ’70s, cities were much more alike than they are today. Since then, their fortunes have diverged, as some cities hollow out while others reach record housing prices and incomes. The Greyhound theory was never “Get to the biggest, densest town,” but “Get to a place that’s thriving.” Autor’s density categories melt the distinctions between San Francisco and Milwaukee, between Silicon Valley and Orlando, and between ski towns and coal country. We know there’s an urban wage premium for entry-level service jobs in San Francisco and Seattle—because it’s the law. Each city has a $15 minimum wage.

That may be how the ladder can be repaired. It’s hard to create and sustain mid-skill work, as cities with manufacturing subsidies have learned. A lot of low-skill work, on the other hand, is “non-tradeable”—that is, hard to outsource or automate. You can’t send a deliveryman’s job overseas, or a janitor’s. It’s up to local governments to make sure those jobs pay more.