Unlocked: The past and future of the city

From Uruk to Tokyo and beyond

I’m on a quick overnight trip to San Francisco, trying to help out some civic minded folks working to improve governance in the Bay Area. This has left me with less time for writing than usual, so I thought it would be a good time to re-run a paywalled post from earlier in the life of Slow Boring, when many of you were not yet members.

This post on cities is broadly connected to the theme of my trip. I believe strongly in the importance of working toward straightforward solutions to our most visible problems, including the obvious point that making it legal to build housing where there is demand for housing makes the economy function better.

But cities are also shaped by much deeper technological forces. You couldn’t have built Los Angeles as we know it in the 18th Century, and you wouldn’t have built Boston as we know it in the 21st century. And that’s what this post is about, taking a very long-term view of urban life — which is a minority of homo sapients’ time on this planet — to try to understand where we may be going. Thanks for your patience while I’m traveling, and I hope you enjoy!

American cities have taken it on the chin over the past few years.

The Covid-19 pandemic and associated business restrictions have disproportionately affected the distinctive amenities of urban life — independent restaurants, theaters, museums, and live events — while the resurgence of violent crime disproportionately threatens their stability. Remote work, meanwhile, poses an immediate threat to the revenue base of core municipalities, which rely on taxing both office buildings and the lunches and after-work drinks of the suburban commuters who populate them.

But there’s a difference between saying that a technological change will result in some bumps in the road that need to be navigated and saying that it will fundamentally derail large urban agglomerations.

And that got me thinking about the longer-run history of cities.

The most up-to-date data source we could find on the sizes of cities over time comes from Stanford’s Ian Morris.

Granting that these are all approximate figures, one big takeaway is that for thousands of years, cities were really small! Back when Memphis, Egypt was the largest city in the world it was much smaller than Memphis, Tennessee is today — and Memphis is not the largest city in Tennessee, and Tennessee is not a particularly large or urbanized state.

But it’s also noteworthy that almost all the cities on this list were the capitals of major empires, or at least what counted as major empires at the time.

The imperial city

One important feature of a city is that the people in it can’t grow their own food; the population is limited by the agricultural production of the lands it controls. Back in the days of Uruk or Babylon, we were talking about pretty small-scale state formation in Mesopotamia.

The vast Roman Empire was much more expansive, and its efficient internal trade routes and supply lines supported a giant metropolis. It’s noteworthy, though, that Rome-the-city sort of collapsed when the Western Roman Empire fell apart. When the first modern census of Rome was undertaken in 1861, the city was home to just under 200,000 people. Today Rome is a big city, just as it was in 600 CE, but the 1,400 years in between were rough.

The Chinese cities on the chart above were also capitals at different times. Kaifeng was the capital of the Northern Song Dynasty and was eventually overrun by the Jin. The Song then established a new capital at Hangzhou, and even though the extent of the Song domain was diminished, it was still large enough to make their new capital the new largest city in the world. Later the Ming Dynasty established Nanjing as the capital, then relocated it to Beijing.

More broadly, it’s also striking how closely tied pre-industrial urbanization was to administration.

When Rome had a million people, even distant Londinium was a somewhat significant city of 30,000-60,000 people at around 100 CE. That’s tiny compared to Classical Rome or modern London but was still a big city compared to Uruk or Thebes from an earlier time. And as Roman authority over Britain waned, the city’s population crashed. It seems like pre-modern economies couldn’t really sustain cities without the city actually ruling over a hinterland that it administered.

Eventually, London became the largest city in the world, and just like its predecessors, it was the capital of a gigantic empire. But what’s interesting about London is that though it’s no longer the biggest city in the world, it’s still a very large city. And its decline has been a purely relative decline. Unlike Rome or Kaifeng or Babylon, it didn’t crash when the empire fell apart. What’s more, London in some sense “wants” to grow. It is very expensive to live in London and were it permitted to tear down more old London buildings and replace them with taller structures and denser residences, there would be a robust market for such dwellings. But this is a fairly new phenomenon.

The global city

London survived the fall of the British Empire by maintaining commercial ties around the world in order to feed itself.

Of course, London is also the administrative center of the UK, and that sustains some of its population. But while shifting the capital from Hangzhou to Nanjing produced a new largest city in the world, nobody thinks that relocating the capital of the UK to Manchester would displace London as Britain’s largest city. Indeed, that’s arguably a reason to make the switch. London is so expensive that land and structures freed up by the departure of government institutions would be swiftly reused by other interests. Northern England, which is cheaper and poorer, would benefit from the jobs.

In the United States, our big cities — New York, Chicago, Los Angeles, Dallas — are really only incidentally centers of administration. None of the four is even a state capital. The older city of San Francisco is the one that hosts California’s Federal Reserve Bank and Federal Circuit Court. Government jobs are still important, but they’re much less important than they’ve historically been to the life of cities.

At the limit, you have Singapore. Both Singapore and Classical Athens are referred to as city-states, but Athens was the administrative hub for a hinterland of farms. Singapore is truly a country that is all city. The green parts of the island are parks; there is no hinterland.

Orchard Road apparently used to have actual orchards and 500,000 pigs lived on the island as recently as 1987, but today Singapore imports more than 90% of its food.

The unusual history of Singapore and its divorce from Malaysia make it an extreme case, but also a good example of the original meaning of the phrase “the exception that proves the rule.” Many big cities still are administrative and logistical hubs for a surrounding hinterland that they are politically connected to, but if that were the essence of the modern city, then Singapore wouldn’t work at all. But Singapore works fine without any political integration — it’s a hub of industrial and commercial activity that trades with the whole world.

The constrained city

All of this is just to underscore how exceptional today’s champion, Tokyo, really is in historical terms. Of course Tokyo is a capital city. And Japan is a large country. But it’s not even close to being the biggest country in the world.

The vastness of the Greater Tokyo Area does not stem from its command of the Japanese hinterland, but from the same forces that keep Singapore fed and kept London afloat as the empire was dismantled: the web of peaceful global commerce that lets people trade efficiently across national boundaries.

But the other thing that makes Tokyo unique is that Japanese public policy encourages Tokyo to grow. It’s much more affordable than London or New York or Los Angeles, with robust housing growth driven by infill-friendly land-use policy. Tokyo adds so much housing that its population keeps growing even as Japan shrinks. It’s all tied together by the incredibly sophisticated rail network of the region, which (as I understand it) features complete integration of intra-urban rail, suburban rail, and intercity rail in a way that is far beyond the capacity of American transportation planners.¹

Lots of other major world cities are much more skeptical of growth. The planning regime in London is so hostile to redevelopment that relative to the preservationist baseline, German bombings during World War II turn out to have had net economic benefits to the city by allowing for higher-density redevelopment.

Matt Grossmann @MattGrossmann

more heavily bombed areas in the Blitz in London exhibit taller buildings today. Without it, a simulation suggests the reduction in agglomeration economies arising from the loss of higher-density redevelopment would lead London’s GDP to drop by some 10%

academic.oup.combillion pound drop: the Blitz and agglomeration economies in LondonAbstract. We exploit locally exogenous variation from the Blitz bombings to quantify the effect of redevelopment frictions and identify agglomeration economies

The People’s Republic of China caps the growth of Shanghai and Beijing based on what the Chinese government characterizes as public health concerns. Some China scholars have told me over the years that the real concern is a super-giant city that could become a single point of failure in case of mass protests. Because the government controls electronic communications, they think they could prevent people from coordinating activities across multiple cities, but the regime can’t afford to let any one city get too important.

The Delhi metro area in India is growing very quickly, and they are also building out their mass transit system at what seems to be a pretty good clip. So I think at some point fairly soon Tokyo will be overtaken, and once again the largest city in the world will be the capital of one of the largest-population countries in the world.

But Tokyo’s reign is a reminder that under modern conditions there’s a fair amount of happenstance to that. It used to be that urban growth was limited by the political and logistical possibilities of obtaining food. Today it’s largely limited by zoning.

Where Americans like to live

In the United States, some cities feature higher incomes than others but are also generally more expensive.

The Hamilton Project charted median annual earnings versus the cost of living and found that there is a clear correlation between the two. But you also have interesting exceptions like Miami and Los Angeles that are very expensive relative to income, presumably because people like good weather and proximity to the beach.

The opposite of Miami is Minneapolis, a high-income city with an average cost of living.

One of my long-time takes from both Slate and Vox is that if you’re footloose and fancy-free you should probably move to Minneapolis. People stay away because it’s cold, but the happiness benefits of good weather occur primarily in the short-term — in other words, you pay a premium for Los Angeles but then your baseline resets and you get used to it. Better to live in Minneapolis and take advantage of the favorable economics to do fun vacations, perhaps to Miami.

But the other thing that strikes me about this chart is that New York is both poorer than D.C. and Boston and also more expensive. If you look at median earnings adjusted for the cost of living, New York is much worse than D.C. or Boston or even Philadelphia.

I think the way to interpret this is that people really enjoy living in the Greater New York City area. Not everyone, of course. But just as you see people voluntarily incurring a real wage penalty to get the weather and beach amenities of Southern California or Florida, you see people with a taste for city living willing to drive their material living conditions down in order to be in New York.

The unwritten future of the American city

As a local homeowner, the super-duper-high real wages in the Greater D.C. area make me kind of worried about remote work. I personally like it here and feel that it’s good for my work. But the numbers suggest that it’s an undesirable place to live. I have previously pegged Seattle as ripe for Zoom-based disruption, and Boston may be resilient due to the dynamics of the labs-and-colleges economy.

But the relatively low wages in New York suggest to me that even if work patterns change, the metro area could actually grow an enormous amount if it wanted to.

There are clearly are people who are (or at least were) tied to the city primarily by commuting needs and who’ll be eager to go live elsewhere. But there actually appear to be more people who like it there but struggle to find a job that pays enough to make it worthwhile. People who right now are doing office jobs in Chicago or Philadelphia or Seattle or D.C. might jump at the chance to do remote work from the NYC area. And I think New York could become really gigantic — à la Tokyo — if the leaders in the city and surrounding states wanted to embrace that future.

Of course, they’d need to change land-use rules significantly and massively upgrade both the physical infrastructure and operational paradigm of their transit system

Yet here again is a case where technology makes things better. It’s always better to have good transportation instead of bad. But to the extent that commuting becomes less important and land occupied currently by offices can be converted to housing, improving the transportation infrastructure becomes less significant.

As for the D.C. area, my hope is that we are underrated rather than doomed and that more people will move here. But I do think it adds up to a strong case that we need more and better-known local amenities.

Maybe our football team getting an actual name will help.

1. Indeed the kind of interlining between metro service and intercity rail that you see in Tokyo, Seoul, and some other Asian cities would be illegal in the United States.