On Friday morning, just before writing this newsletter, I spoke to a conference of the Regional Science Association in Alicante, Spain. Notice the choice of preposition: “to,” not “at.” Given family commitments, I couldn’t attend in person, so this was a Zoom presentation from home — something we’ve all become familiar with over the past three years.
The title of my session was “Is the New Economic Geography Still Alive and Well in 2023?” The obvious subtext was, is it still alive and well given our ability to do what I was doing — participate in a discussion with people thousands of miles away?
Some background: Economic geography — the study of where people do stuff and why — has been around for a very long time. The “new” economic geography refers to a particular way of approaching the subject that makes as much use as possible of buttoned-down, formal economic models. My own most cited academic paper, “Increasing returns and economic geography,” published in 1991, was one of the early works in this genre.
What’s the point of formal modeling here? It helps clarify one’s thinking, and quite often yields insights that should have been obvious — that can, once arrived at, be expressed easily in plain English — but that you didn’t have before. Maybe the most important insight from the N.E.G. is that there is always a tension between the forces of agglomeration, which tend to make activity clump together, and centrifugal forces that tend to make it spread out. And changes in technology can push the economy to a tipping point in either direction, transforming where we work and live.
Thus in the second half of the 19th century the rise of large-scale manufacturing production, which gave industry an incentive to concentrate near the biggest markets, and railroads, which made it easy for distant farmers to feed urban populations, led to the rise of the “manufacturing belt,” a concentration of industry in the Northeast and the inner Midwest.
Sometime around the 1920s, however, centrifugal forces gained the upper hand. Trucking made it less important to be near a rail hub, while electrification led to a redesign of factories — hulking mills replaced with sprawling one-story structures, best not located in dense urban areas. So industry and wealth spread out. The manufacturing belt gradually dissolved, and in general, disparities in income across the United States narrowed over time.
In a way, then, U.S. economic geography became less interesting. And to be honest, the early N.E.G. literature had a somewhat backward-looking, almost steampunk feel, dwelling on things like the rise of the manufacturing belt and the emergence of local manufacturing clusters like the detachable collar and cuff industry of Troy, N.Y. There was also considerable interest in China, whose rise as a manufacturing superpower has been accompanied by the growth of industrial clusters reminiscent of those prominent in America circa 1900.
But wait, the story isn’t over.
Around 1980 the forces driving economic geography changed again, with the rise of a knowledge economy. Technology-intensive firms, whether or not they were explicitly in the technology sector, wanted access to a large pool of highly educated workers, which by and large meant locating in big coastal metropolitan areas. Growing employment opportunities in these areas drew in even more highly educated workers, so regional disparities began to widen again.
You can see the reversal of fortune in a couple of charts. Here’s the ratio of per capita income in Mississippi, which has consistently been one of America’s poorest, less educated states, to income in Massachusetts, which has always been near the top:
The gap between these states narrowed a lot between the 1920s and 1980; the American economy was looking increasingly like a level playing field, at least in terms of location. But this convergence went into reverse after 1980.
A study from the Brookings Institution went at this issue from a different perspective, looking at the dominance of a handful of “superstar” cities. It showed these cities rapidly pulling away from the rest of the country:
And while big coastal metros were booming, large parts of the United States — especially what some economists have called the “eastern heartland” — were left behind.
Then came Covid.
For a few months in 2020 it looked as if population density itself might be a major source of contagion risk. Once we learned more about the disease, however, and especially after vaccines became available, this concern vanished.
But Covid did something else. Remote work has been technologically possible for some time, ever since large numbers of people gained access to high-speed internet. But the possibility went largely unused until fear of contagion forced much work to become remote. Once that happened remote work reached critical mass: Millions of workers learned how to interact over the internet, and found their co-workers doing the same thing. And by and large they liked it, for obvious reasons: absence of commuting, easier management of work-life balance, and so on.
As a result, even though in many ways life has returned to its prepandemic normal, many people are still working from home and office occupancy remains depressed:
So does this mark the beginning of a sustained decline in superstar cities? I don’t think so.
While remote work is pretty clearly here to stay, there’s a big difference between fully remote work, in which you never visit the office, and hybrid work, which involves working from home only two or three days a week.
Fully remote work can be done from anywhere, and you might expect a significant number of fully remote workers to relocate to smaller cities that offer urban amenities like walkable downtowns. Early indications are, however, that fully remote work will remain a fairly small niche of the work force. From what I’ve seen, both employers and workers are increasingly seeing the value of the informal interactions that come when you go into the office, at least some of the time.
And hybrid workers still need to live within a major metro area’s commuting zone.
In fact, if anything, the rise of hybrid work may reinforce the advantages of superstar metros — or perhaps more accurately reduce one of their major disadvantages. After all, one of the major hassles of working in someplace like New York City is the time and expense of commuting. Hybrid workers don’t have to commute as much as full-time in-office workers. Alternatively, they can move farther away from their jobs, getting cheaper housing, while commuting longer distances, but on fewer days.
So will new technology remove the edge superstar metros have gained over the past generation? Probably not. It may force big changes in their internal structure: What will we do with all that excess office space? But at this point, I don’t see us going through another historical reversal in the economy’s geography.