Big city, higher pay

Say you’re a software developer in your hometown of Logan, Utah, a scenic college town of 155,000, earning $41 an hour. How might you increase your wage?

If you look up what software developers in Los Angeles are making, you might think moving could do the trick: The median hourly wage for software developers in LA is 60 percent higher than in Logan. Figure 1 offers a snapshot of this big-city wage premium for four different occupations in two small cities and two large ones. It’s a pattern that has been well documented across the U.S. as well as in other countries.

“Firms definitely find a reason to pay workers more when they’re in big cities,” Institute visiting scholar and Columbia Business School professor Conor Walsh said. And it’s not just that wage levels are higher in big cities, Walsh explains. “People who work in big cities tend to have steeper wage experience profiles—that is, their wages rise faster with experience than comparable workers in small cities.” This is true even for workers with similar educational backgrounds, occupations, and years of experience.

So why do wages grow faster in big cities? It’s not something in the air or the water, of course. But is there something about the way economic activity is organized that makes workers in big cities more productive, causing their wages to grow more speedily? Or are more productive workers the ones who choose to move to big cities?

Randomly assigning workers to a big city

Determining whether cities make workers more productive or whether more productive workers choose to live in cities is tricky.

“In the ideal experiment, what you’d love to be able to do is take a bunch of people and randomize their locations,” said Walsh, who is a co-author on two Institute working papers that investigate the sources of the big-city wage premium. Put some people in a major metropolis, put others in small towns, let them work, and observe how their wages evolve over time.

Economists aren’t allowed to run experiments like that, but governments sometimes can. The Danish government decided in 1986 that it would assign refugees where to live so that municipalities across the country would all share in integrating the new arrivals. The council that made the assignments had access to only a few pieces of information about each refugee: date of birth, marital status, number of children, and nationality. The council did not see education or employment history, so refugees with different educational and occupational backgrounds were distributed randomly across the country.

Denmark has a population of about 5.9 million people, and about 2 million of them live within commuting distance of Copenhagen, the country’s one big city. (Aarhus, the country’s second-largest city, has about 300,000 people.) Walsh and his co-authors, former visiting scholar Fabian Eckert and Mads Hejlesen, compare the wage trajectories of refugees assigned to Copenhagen to those of refugees assigned elsewhere. They limit their sample to male refugees between the ages of 19 and 55 who arrived in Denmark between 1986 and 1998 from eight originating countries, for a total of about 20,500 individuals. Limiting the sample in this way keeps the comparison among people whose experiences in the labor market are likely to be similar.

So what happened? Initially, wages of refugees assigned to Copenhagen and those assigned elsewhere were essentially the same. But then they diverged, with wages in Copenhagen growing 35 percent more each year than wages of refugees living elsewhere who had a similar number of years of labor market experience.

So cities are making workers more productive. But how? “Over time, a refugee assigned to Copenhagen is increasingly likely to work at more productive firms and in more skill-intensive occupations and industries than one settled outside Copenhagen,” the economists write. After 15 years in Denmark, 40 percent of refugees assigned to Copenhagen were working in high-skill occupations (such as professional or technical activities), compared with only 27 percent of workers assigned elsewhere. Outside of Copenhagen, refugees were more likely to transition to manual industries, such as manufacturing, where wages aren’t as high.

The dramatic rise of business services firms

Walsh, Eckert, and Hejlesen’s findings suggest the source of big-city wage growth stems from the types of firms and occupations found in big cities. “It really seems to be about the distribution of firm types between cities just differing radically,” said Walsh. “In Copenhagen, there’s a much higher proportion of high-productivity establishments that pay higher wages, on average.” It’s a pattern that holds in major cities across Europe and the United States.

What are these highly productive, high-paying firms? This is the subject of a second Institute working paper by Walsh and Eckert, with co-author Sharat Ganapati, that investigates why wages in big cities have outpaced wages elsewhere.

The economists find that almost all of the big-city wage growth in the U.S. since 1980 comes from a single sector: business services. Economists often use the Census Bureau’s industry classification system, called NAICS, which groups together industries that use similar production methods. Business services encompasses the information sector (e.g., telecommunications, data processing, web search portals); professional, scientific, and technical services (e.g., legal services, architecture, engineering, veterinary services); finance and insurance; and management of companies and enterprises.

Figure 2 shows what wages looked like in 1980 and 2015 across commuting zones, which are ordered according to the density of their population. In 1980, business services workers in the densest cities earned 42 percent more, on average, than business services workers in the least dense areas. In 2015, business services workers in the densest cities earned 111 percent more. Workers in other industries, meanwhile, do earn more in big cities, on average, but not much more—and the gap between big cities and elsewhere hasn’t changed in this 35-year span.

A location mystery

Another well-documented trend in wages is the growth of the college wage premium—for decades, the gap between average wages of those with a college degree and those without has grown. But intriguingly, the wage growth of big-city business services workers is not primarily a story about education. While the business services workforce does tend to be college-educated, the fraction of employees in this sector with a college degree has not changed much over time. And workers in the business services sector without a college degree have experienced faster wage growth than workers elsewhere, too. Meanwhile, outside of the business services sector, neither workers with or without a college degree have experienced more wage growth in big cities than elsewhere.

It’s also not the case that the business services workforce has been growing rapidly. About 15 percent of the U.S. workers were employed in business services in 1980, Walsh said, and that number has stayed relatively stable since about 1980. “What has changed a lot,” Walsh said, “is payroll. Depending on the dataset you look at, [the payroll of business services firms] used to account for about 20 percent of total U.S. payroll. Now it’s closer to 40 percent.”

Identifying the sector that is responsible for big-city wage growth only deepens the mystery, in a way. After all, many business services firms could locate just about anywhere. Software companies don’t need to be in San Francisco to create software. Biomedical startups don’t need to be in Boston to make medical devices. That highly productive firms choose the largest metropolises suggests “there must be some productive advantage to doing so,” Walsh said.

That productive advantage, the economists believe, is tied to information and communication technologies (ICT)—mainly, computer hardware and software. While many sectors use ICT, the business services sector adopted ICT earlier and in larger quantities than other sectors. Then, between 1980 and 2000 in particular, the price of ICT dropped, meaning the price of a major input in this sector kept getting cheaper and cheaper.

Why does this mean wages at business services firms grew so fast in big cities specifically? “Software investments tend to be very scalable technologies,” Walsh explained. “If you buy a piece of custom software for your organization, you only need to buy it once, and you can roll it out to everyone who works with the software at your firm. Larger firms can afford to pay large custom software fees.” As a result, large firms not only spend more on ICT—they spend more per employee on ICT than small firms do.

And business services firms in big cities are larger than business services firms elsewhere. That combination of size plus intensive use of ICT that business services firms in big cities enjoy gave their workers an incredible advantage in productivity. And when workers’ productivity goes up, so do their wages.

Walsh points to Goldman Sachs as one example of this process at work. “Goldman Sachs uses a huge amount of IT for various things, right—for trading, but also for research purposes. And as IT gets cheaper, the incredibly talented not just bankers but software engineers that work for Goldman Sachs see large increases in productivity.”

In sum: Big cities are home to many large business services firms. Large business services firms make use of a lot of information and communication technology, more than smaller firms or firms in other sectors. The price of that technology has dropped over time, meaning the firms can buy more and more of it—becoming more and more productive in the process. The result is faster wage growth.

The ultra-metropolis of our future?

If moving to a big city leads to faster wage growth for workers, why doesn’t everyone move to the city, resulting in one ultra-metropolis? Well, because density creates downsides, too. There’s traffic, which is not just annoying but harmful to mental health. Air pollution. Crowded restaurants, stores, parks. Less nature. And, as most city dwellers know all too well, there’s the high cost of housing. New research from Institute advisor Jesse Rothstein, David Card, and Moises Yi that explores how and why wages vary across places also considers how local housing costs compare with local wages. The economists find that higher housing costs in larger cities more than completely offset the higher wages.

“You could think of different groups of people sorting into different cities based on how much productivity advantage they get versus how much they don’t like congestion,” Walsh said. For some, the glitter of New York City will always shine brightly. For others, not so much.

“There may just not be enough space for more of this process to continue,” Walsh said. The pandemic and post-pandemic years were tumultuous for the populations of big cities. Just where people will end up choosing to live and work remains to be seen.

Landon Peterson provided research and editorial assistance for this article.