Wednesday, March 17, 2021

Will Workers Disperse from Cities?

Predictions that technology shifts will cause urban job concentrations to disperse have been made a number of times in the last half-century or so. The predictions always sound plausible. But up until the pandemic, the predictions kept not happening.

Here's an example from a 1995 book City of Bits, by an MIT professor of architecture named William J. Mitchell. He wrote a quarter-century ago, while also making references to predictions a quarter-century before that (footnotes omitted): 

As information work has grown in volume and importance, and as increasingly efficient transportation and communication systems have allowed separation of offices from warehouses and factories, office buildings at high-priced central business district (CBD) locations have evolved into slick-skinned, air-conditioned, elevator-serviced towers. These architecturally represent the power and prestige of information-work organizations (banks, insurance companies, corporate headquarters of business and industrial organizations, government bureaucracies, law, accounting, and architectural firms, and so on) much as a grand, rusticated palazzo represented the importance of a great Roman, Florentine, or Sienese family. ... 

From this follows a familiar, widely replicated, larger urban pattern--one that you can see (with some local variants) from London to Chicago to Tokyo. The towers cluster densely at the most central, accessible locations in transportation networks. Office workers live in the lower-density suburban periphery and commute daily to and from their work.  ... 

The bonding agent that has held this whole intricate structure together (at every level, from that of the individual office cubicle to that of CBDs and commuter rail networks) is the need for face-to-face contact with coworkers and clients, for close proximity to expensive information-processing equipment, and for access to information held at the central location and available only there. But the development of inexpensive, widely distributed computational capacity and of pervasive, increasingly sophisticated telecommunications systems has greatly weakened the adhesive power of these former imperatives, so that chunks of the old structure have begun to break away and then to stick together again in new sorts of aggregations. We have seen the emergence of telecommuting, "the partial or total substitution of telecommunication, with or without the assistance of computers, for the twice-daily commute to/from work."

Gobs of "back office" work can, for example, be excised from downtown towers and shifted to less expensive suburban or exurban locations, from which locally housed workers remain in close electronic contact with the now smaller but still central and visible head offices. These satellite offices may even be transferred to other towns or to offshore locations where labor is cheaper. (Next time you pay your credit card bill or order something from a mail-order catalogue, take a look at the mailing address. You'll find that the envelope doesn't go to a downtown location in a major city, but more likely to an obscure location in the heartland of the country.) 

The bedroom communities that have grown up around major urban centers also provide opportunities for establishing telecommuting centers small, Main Street office complexes with telecommunications links to central offices of large corporations or government departments. As a consequence, commuting patterns and service locations also begin to change; a worker might bicycle to a suburban satellite office cluster or telecommuting center, for example, rather than commute by car or public transportation to a
downtown headquarters. Another strategy is to create resort offices, where groups can retreat for a time to work on special projects requiring sustained concentration or higher intellectual productivity, yet retain electronic access to the information resources of the head office. This idea has interested Japanese corporations, and prototypes have been constructed at locations such as the Aso resort area near Kumamoto ...

More radically, much information work that was traditionally done at city-center locations can potentially be shifted back to network-connected, computer-equipped, suburban or even rural homes. Way back in the 1960s, well before the birth of the personal computer, James Martin and Adrian R. D. Norman could see this coming. They suggested that "we may see a return to cottage industry, with the spinning wheel replaced by the computer terminal" and that "in the future some companies may have almost no offices." The OPEC oil crisis of 1973 motivated some serious study of the economics of home-based telecommuting. Then the strategy was heavily promoted by pop futurologists of the Reaganite eighties, who argued that it would save workers the time and cost of commuting while also saving employers the cost of space and other overhead. The federal Clean Air Act amendments of 1990, which required many businesses with a hundred or more employees to reduce the use of cars for commuting, provided further impetus. ...

In the 1960s and early 1 970s, as the telecommunications revolution was rapidly gaining momentum, some urbanists leaped to the conclusion that downtowns would soon dissolve as these new arrangements took hold. Melvin Webber, for example, predicted: "For the first time in history, it might be possible to locate on a mountain top and to maintain intimate, real-time and realistic contact with business or other associates. All persons tapped into the global communications net would have ties approximating those used today in a given metropolitan region." ...

But the prophets of urban dissolution underestimated the inertia of existing patterns, and the reality that has evolved in the 1980s and 1990s is certainly more complex than they imagined. The changing relative costs of telecommunication and transportation have indeed begun to affect the location of office work. But weakening of the glue that once firmly held office downtowns together turns out to permit rather than determine dispersal; the workings of labor and capital markets and the effects of special local conditions often end up shaping the locational patterns that actually emerge from the shakeup.
I love the passage in part because it starts of in the first paragraph talking about how dense central business districts "represent the power and prestige of information-work organizations," which makes it sound as if downtown urban areas are nothing but an ego trip for top executives, but then ends with some comments about how economic factors "labor and capital markets" actually end up shaping the results. 

The economic patterns of big cities have changed. I have discussed "How Cities Stopped Being Ladders of Opportunity" (January 19, 2021), because in recent decades they have been places where the more-educated could earn higher wages, but they have stopped being places where the less-educated could earn higher wages. 

But moreover, when Mitchell in his 1995 book referred to "the need for face-to-face contact with coworkers and clients," he was seeing only part of the picture. Yes, contact with coworkers and clients within a firm matters, but it's also true that firms of a certain type often bunch together geographically. It seems important to be geographically located near workers and clients from other firms, too. I've written a bit about this "economics of density," and offer some links, in "Cities as Economic Engines: Is Lower Density in Our Future" (August 14, 2020). 
Hannah Rubinton offers another piece of evidence in "Business Dynamism and City Size" (Economic Synopses: Federal Reserve Bank of St. Louis, 2021, Number 4). The points represent data for individual cities. The horizontal axis shows the population of the city. The vertical axis of the top panel shows the "establishment entry rate," which is the rate at which new business establishments are started in a city. An "establishment" includes both a new business or a new location for part of an existing firm. In the bottom panel, the vertical axis shows the "establishment exit rate." The payoff for these figures is that if you plot the data for 1982, you can that larger cities tended to have lower rates of entry and exit (the solid lines slope down), but by 2018 the larger cities tended to have higher rates of entry and exit (the dashed lines slope up.)

This pattern reflects that in the last few decades, a substantial part of economic dynamism, productivity growth, and wage growth has been happening in the larger cities. As Rubinton notes: 
At the same time, large and small cities have diverged on several important dimensions: Large cities increasingly have a more educated workforce and offer higher wage premiums for skilled workers. Given that dynamism is important for productivity and economic growth, the differential changes in dynamism across cities could be important to understanding the divergence in wages and skill-composition between large and small cities. ... [T]hese patterns are consistent with competition becoming tougher in large cities relative to small cities. Large cities have become more congested than they were in 1980: As population has grown and technology has improved, rents and wages have increased. Less-productive firms that cannot afford the higher prices are more likely to exit, leaving room for new firms to enter.
Maybe the aftereffects of the pandemic will change all this. I tend to believe that some of the shift to telecommuting in this last year will persist. But I'm also very aware that predictions about how jobs "can potentially be shifted back to network-connected, computer-equipped, suburban or even rural homes" have been around for decades. Yet downtown business districts and other clusters of economic activity continue to persist and grow, which suggests strong underlying economic forces at work.