[Over the course of this post I pontificate about things that I only have partial knowledge of. I think what I have to say is valuable because it creates thoughts that can be tested. However, equally valuable is this disclaimer.]
One thing that has always perplexed me is the lack or correlation between the size of a given city and how well it is managed. To illustrate why this perplexes me, if a business does a bad job attracting customers, then that business will go under. If you then extend that interpretation to include cities, and think of the residents of that city as the city’s “customers”, then you would think that the cities that do the best job managing its resources and population, would have the highest population, and vice versa. But that doesn’t seem to be the case. Millions live in LA, despite it’s bad air quality, traffic, and a homelessness crisis. Millions more flock to NYC, despite its lackluster Subway quality, even worse traffic (by some measures), and (a personal pet peeve) a lack of public restrooms.
You might think that my lamentations of NYC and LA are unfair and inaccurate. There are certainly things that LA and NYC do right. And you might think that the correlation instead runs negative. Meaning that large cities are harder to manage, so they are inherently more likely to have more problems. That seems to be a general way people think about where these problems of cities come from. I disagree with this interpretation because if you look at one of the world’s largest cities, Tokyo, you see the opposite thing happening. Tokyo is both very large, and very low in crime, having a robust and successful economy (relatively speaking), and boasting one of the worlds best transit systems. (Side note: it was my first time going to Tokyo that got me interested in urban design to begin with.) And you can also find examples of very small cities and towns with a very poor and destitute population. It just doesn’t make sense. And I’m not sure what the actual answer is, but I have a few hypotheses.
It’s possible that this is merely a statistical phenomenon. The same thing can be observed in colleges. At one point the Bill and Melinda Gates Foundation at one point believed that smaller colleges were more successful than larger ones. (Citation: hearsay.) However, upon further research of the data, the true reason was found. It turns out that this simply has to do with the fact that there are more smaller colleges than larger ones. And so, simply by random chance, you are going to find more good colleges that are small than that are large. If you were to fill a bucket with 110 colleges, with their rating of “goodness”. And let’s say 10 of them are large, 100 small. You could have a situation where larger colleges are more likely to be successful, but you are still able to find more examples of successful small colleges. Let’s say 40% of large colleges are “good”, and 20% of small colleges. Then you would end up with a situation where you are able to find 4 examples of good large colleges, and 20 examples of good small colleges. And come to the conclusion that small colleges are better. Because 20 is larger than 4. (A similar situation can be observed in measuring the effectiveness of the COVID-19 vaccine. In places with a particularly high rate of vaccination, you end with more breakthrough cases than unvaccinated cases. That does not prove the vaccine is ineffective. But I digress.) And those 20 good colleges don’t even have to be particularly good. Simply by random chance, smaller colleges are more likely to admit and graduate successful students because they have fewer opportunities for bad students to mess up the average. This is true for the same reason that if you take a jar of jellybeans, the smaller the handful, the more likely you are to take more of the same color. And the larger the handful, the more likely you are to grab a distribution of colors that represents the jar as whole. So if I were to blindly grab 3 jellybeans from the jar, and all 3 of them are red, that wouldn’t prove that I’m really good at grabbing red jellybeans and really bad at grabbing purple ones. The same thing could be true for colleges, cities, companies, etc.
This type of phenomenon could explain some of the desire for small townness in popular culture. Larger cities have more opportunities to experience issues with homelessness, because there are more people, and more people means more homeless people, for example. And the same thought process could be true for other aspects of city management. Personally, I don’t buy this explanation. It does somewhat explain why there are a wide variety of socioeconomic situations in small settlements, but it doesn’t stack against my Tokyo counterexample. Barring even that, it does beg the question as to why these small successful cities don’t inevitably grow into large successful cities. (The same question arises with colleges, but this post is already running long.)
Here we get to some of the structural changes that might be impeding the success of various regions. I’ve been reading a book called, “Cities and Regions as Self Organizing Systems” by Peter Allen, and they do touch on this topic. I’ll be going over my take on it plus some additional perspectives. And in my opinion none of these explanations provide a perfectly complete picture of what is going on, but they create hypotheses that can be tested and built on.
To start with, the area that encompasses the official boundary of a given city is usually much smaller than the metropolitan area we normally associate with it. I live (at the time of writing) in Madison, WI. The population of Madison is about 250,000 if you only include the city proper, but about 500,000 if you include the surrounding suburbs. If you look at a map of the city’s limits you would see one of the most jagged areas imaginable. You could easily mistake it for a gerrymandered district. Filling the gaps is an amalgamation of towns that have either refused to integrate themselves into the city of Madison, Or have otherwise made some kind of deal with the city. Whether or not they have done this, it is often the case that property owners (in Madison and elsewhere in America) will fight tooth and nail against high density development in their neighborhood. Whether it be by forming their own section of a given city that doesn’t need to abide by the same rules, their own township, or simply protesting at council hearings. Meaning that even within one of these given “cities,” it is hard to expand on the population because (though theoretically possible to build) there is no more housing to put people in. So, without the ability to expand outward, and without the ability to expand inward and upward, the success of a city can be measured by how effectively they can segregate or shoe away “bad” residents and attract “good” residents. Which is easier to do if your town is small. But it does create this dichotomy of winners and losers in the economic game of city quality.
But this does beg the question in my mind. Surely there are at least some examples of these cities successfully integrating the less well off and/or creating a high density environment that is desirable to the wealthy. You could say that Manhattan is an example of the latter. And it’s not like every non-affluent place is also a terrible place to be. However, this explanation doesn’t quite do it for me. In the case of Manhattan, there’s a little bit of a chicken and egg problem. NYC has always been a large city compared to it’s American counterparts. This is despite many of the issues that it has faced throughout its history. It seems more likely to me that property values on the island skyrocketed for other reasons (more on that in a bit) and got it to the point where the property values would remain high regardless of how much development took place. And it’s not like the city is fully in favor of development. It’s learned ways to compromise on these things, with setback requirements, limits on developing historical landmarks, small commercial spaces at the base of their skyscrapers, and things like that that make the area attractive (and expensive) for residents and tourists alike, driving up property values. So the marginally higher population of NYC can partially be explained by their marginally better handling of high density. However, given the existence of similarly large populations with completely different ways of dealing with this issue. I don’t think it tells the whole story. And if that were the case, how come we don’t see those aforementioned small towns growing into amazing metropolises?
All in all, this is a pretty weak argument in my opinion. Worth exploring because it reveals other issues with the way that cities are designed in America. But it doesn’t say anything about the consistent rule of city size distribution. What I just destined is a contemporaneous problem that mostly takes place in America. Other parts of the world do not necessarily have quite the same attitude about urban planning. At other times in history America didn’t have the same attitude about urban planning. Yet this rule has remained true throughout region and time. We cannot look at contemporary explanations to find our answer. And we shall search, instead for things that are more consistent.
It’s possible that part of the answer lies in the common explanation for why cities form in the first place: location. Cities are (mostly) not artificial entities. They come out of the existence of settlements of individuals exploiting resources, strategic positions along trading routes, etc.. And that seems to say more about a city’s size than anything else. And this is where our modelers in “Cities and Regions as Self Organizing Systems” come in. They created simplified simulations of population dynamics. The mathematics they use is a little bit over my head, but what I can tell you is that these simulations tested differing population distributions based on random events. Trying to take into consideration factors such as the natural carrying capacity, transportation costs, innovation that then increases the carrying capacity of a given location, and so on. One thing they discovered, perhaps also not surprisingly, is that there are several stable positions that can come out of the distribution of a population. If you run the simulation several times, you will get several arrangements that will then remain consistent indefinitely. This demonstrates, perhaps, that there are some things (that may or may not be random and unknowable) that can happen early in urban development and have a lasting effect for generations.
This is observed in the real world. The proportion of the largest cities in the world has stayed consistent with itself for most of human history. (Barring instances where new lands are discovered. Such as America.) This model suggests that it was not necessarily inevitable that London would be the largest city in Europe. That if you were to run history again and allow the butterfly effect to take hold, then some other city might have just as well taken that spot and kept it. Perhaps it could have just as easily have been Rotterdam or Southampton to become the largest city in Europe instead of London. Both are located strategically. You might argue that London is located on the Thames, and that is what makes it so large. Then I ask you Why London is not located further up the Thames where it is more centrally accessible to the rest of the nation like Paris is. Conversely, why isn’t Paris located closer to the shore? You might say that this is because they are both capital cities. But it is not inevitable that the capital is the largest city in the country. The answer might just be random chance.
This consolidation of location is further exacerbated by the existence of infrastructure feeding into and out of these population centers. As well as maintaining them. Once a highway is built to Chicago, it becomes difficult to convince businesses to move their cargo into another city where it will be harder to get to. Once you have the infrastructure to feed cargo, you need to build infrastructure for the water and sewage system, this attracts more businesses, which attracts more people, and the cycle continues.
However, if a municipality fails to maintain the infrastructure necessary to maintain the population, then in theory, over a long enough period of time, an equally strategic location could snatch at the opportunity to start to build its own infrastructure. So maybe the answer is that how well managed a city is DOES have an effect on it’s population, just on a multigenerational timescale. And perhaps the issue of randomness is what keeps the proportion of the largest cities the way that they are. If you imagine large cities and small cities on an infinite timescale, some of the small cities could, in theory, overtake the larger ones eventually. But just as randomness causes them to be good in the first place. Randomness will cause them to be bad in the future, while the large cities will remain steady, with location and infrastructure to back it.
Earlier in this essay I said that the distribution of the largest cities has mostly been consistent over the course of history. Mostly. There was a time when Tenochtitlan of the Aztec empire was one of the largest cities in the world before the Spanish arrived. That empire has fallen, and the city has come with it. Maybe there is something to be said for that. Maybe over the course of generations, there is some significance to the size of the city and its ability to maintain its population. But this all needs further research.