If you have a long, painful, soul-crushing commute, Tesla has a solution for you: Full Self-Driving (their autonomous, but still supervised, self-driving technology). And it makes sense that Tesla would position its product in this way. A great deal of our built environment (the vast majority of it in some geographies) has been designed around the car. We are dependent. And this is an obvious solution to its negatives.
To be clear, I'm excited about autonomy, which is why it's a frequent topic on this blog. But the urbanist in me can't help but think that positioning it in this way is in some ways a solution to the wrong problem. Here's an alternative solution: live and work in a walkable, transit-oriented community.
Imagine, for instance, pitching this Tesla positioning to a Tokyoite. Tokyo is reported to have the highest railway modal split in the world. According to some measurements, only something like 12% of trips in the city are done by car. So if you said, "FSD is the solution to your long and boring commute. Now you can just sit, relax, read a book, do work, or play on your phone!" it wouldn't be a stretch to imagine Tokyoites saying that they already do this on a train.
Of course, Tokyo is a unique place, and there are lots of car-dependent cities where there is simply no other practical option. I also recognize that housing attainability is a major driver of sprawl. In these cases, FSD represents a meaningful quality-of-life upgrade.
Again, I support this happening, but at the same time, I worry about it placating us into thinking that we've solved one of the major negatives of urban sprawl. Yes, you have to sit in a car for two hours each day, but now you're not actually driving. Isn't that, like, so much better? In a best-case scenario, we maintain the status quo when it comes to our built environment. And in the worst-case scenario, it leads to even more sprawl.
This is an open question that we have on this blog: To what extent will self-driving cars increase our willingness to commute? Historically, new mobility technologies have promoted urban sprawl because they allowed us to travel greater distances in the same amount of time. Consider streetcar suburbs and then our car-oriented suburbs.
A big part of the AV argument is not that they will solve traffic congestion (they won't); it's that they will make your commute suck a lot less, and in an even rosier scenario, become a kind of "third space" where people work, relax, or whatever. This, in turn, will make sprawl more widely palatable.
But the more I think about this, the less I believe it. Marchetti's Constant tells us that humans have generally maintained a consistent "time budget" for commuting irrespective of the technology being used. Will this time really be different?
On the flip side, there are many who would argue that urban sprawl is a natural market outcome. Not everyone wants the "utopian, socially-engineered dream" that urbanists and YIMBYs like me want. And this is a fair response. I believe in individual freedoms. Give people housing options (we're very bad at this) and let them choose where they want to live.
But we should acknowledge the tradeoffs. Traffic congestion is a clear byproduct of urban sprawl and land-use patterns that leave no other practical option for getting around. Complaining about traffic is complaining about sprawl. One more lane or cars that drive themselves have not been shown to change this relationship.
Sprawl also contributes to greater loneliness and declines in happiness. In 2000, Robert Putnam argued in his book, Bowling Alone: The Collapse and Revival of American Community, that, roughly speaking, every 10 minutes of additional travel time leads to a 10% reduction in social connections. We spend less time with our families, friends, and communities.
There's little doubt that self-driving cars will make commutes more tolerable. But perhaps that's not ambitious enough.
Sprawl is how much of the US provides new housing, and so it's interesting to ask the opposite question: Which cities are actually building new housing in walkable neighborhoods? Here is a study published this week by the Terner Center for Housing Innovation at UC Berkeley that looked at exactly this. What they did was divide all US neighborhoods into five categories based on vehicle miles traveled (VMT) per resident in 2023.
The categories:
Very Low VMT - 12 miles per person per day
Low VMT - 17.3 miles per person per day
Mid VMT - 21 miles per person per day
High VMT - 25.5 miles per person per day
Very High VMT - 37.5 miles per person day
These seem like oddly specific distances, but it's what they used to sort new housing supply. Here's all of the US:
Urban sprawl is how much of the US provides new housing. And here's Conor Dougherty in the New York Times arguing that America needs more of it to fix its housing shortage:
Even if all the regulatory restraints were removed tomorrow, developers couldn’t find enough land to satisfy America’s housing needs inside established areas. Consequently, much of the nation’s housing growth has moved to states in the South and Southwest, where a surplus of open land and willingness to sprawl has turned the Sun Belt into a kind of national sponge that sops up housing demand from higher-cost cities. The largest metro areas there have about 20 percent of the nation’s population, but over the past five years they have built 42 percent of the nation’s new single-family homes, according to a recent report by Cullum Clark, an economist at the George W. Bush Institute, a research center in Dallas.
The obvious benefit is that the resulting housing tends to be cheap. The above article is filled with examples of people buying large homes for a few hundred thousand dollars in newly formed communities across Texas. And if you live in a high-cost city, the social algorithms have almost certainly found you at some point with a shockingly cheap house in one of these places. But, Dougherty also admits that sometimes this may be the only redeeming quality:
Escobar told me he moved to Princeton because he could find a big house there for less than $300,000, but now the city is home, and he didn’t like where it was headed. Over the next four years, he said, his goal is to redevelop the downtown, try to attract offices where locals can work and build out a park system that voters recently funded with a bond measure. “You ask anybody what they love about Princeton, and it’s simply just the affordability,” Escobar told me. “We need to be more than that.”
According to the article, this isn't necessarily a problem, because it's just how cities are built in this day and age. What you do is start with low-cost housing in fringe locations. You grow as quickly as possible until traffic becomes "godawful" and vital infrastructure can't keep up. Then you implement moratoriums on new housing, and start working on other uses like, you know, employment. Eventually, after all this chaos is complete, you end up with something that possibly resembles a real city.
If you have a long, painful, soul-crushing commute, Tesla has a solution for you: Full Self-Driving (their autonomous, but still supervised, self-driving technology). And it makes sense that Tesla would position its product in this way. A great deal of our built environment (the vast majority of it in some geographies) has been designed around the car. We are dependent. And this is an obvious solution to its negatives.
To be clear, I'm excited about autonomy, which is why it's a frequent topic on this blog. But the urbanist in me can't help but think that positioning it in this way is in some ways a solution to the wrong problem. Here's an alternative solution: live and work in a walkable, transit-oriented community.
Imagine, for instance, pitching this Tesla positioning to a Tokyoite. Tokyo is reported to have the highest railway modal split in the world. According to some measurements, only something like 12% of trips in the city are done by car. So if you said, "FSD is the solution to your long and boring commute. Now you can just sit, relax, read a book, do work, or play on your phone!" it wouldn't be a stretch to imagine Tokyoites saying that they already do this on a train.
Of course, Tokyo is a unique place, and there are lots of car-dependent cities where there is simply no other practical option. I also recognize that housing attainability is a major driver of sprawl. In these cases, FSD represents a meaningful quality-of-life upgrade.
Again, I support this happening, but at the same time, I worry about it placating us into thinking that we've solved one of the major negatives of urban sprawl. Yes, you have to sit in a car for two hours each day, but now you're not actually driving. Isn't that, like, so much better? In a best-case scenario, we maintain the status quo when it comes to our built environment. And in the worst-case scenario, it leads to even more sprawl.
This is an open question that we have on this blog: To what extent will self-driving cars increase our willingness to commute? Historically, new mobility technologies have promoted urban sprawl because they allowed us to travel greater distances in the same amount of time. Consider streetcar suburbs and then our car-oriented suburbs.
A big part of the AV argument is not that they will solve traffic congestion (they won't); it's that they will make your commute suck a lot less, and in an even rosier scenario, become a kind of "third space" where people work, relax, or whatever. This, in turn, will make sprawl more widely palatable.
But the more I think about this, the less I believe it. Marchetti's Constant tells us that humans have generally maintained a consistent "time budget" for commuting irrespective of the technology being used. Will this time really be different?
On the flip side, there are many who would argue that urban sprawl is a natural market outcome. Not everyone wants the "utopian, socially-engineered dream" that urbanists and YIMBYs like me want. And this is a fair response. I believe in individual freedoms. Give people housing options (we're very bad at this) and let them choose where they want to live.
But we should acknowledge the tradeoffs. Traffic congestion is a clear byproduct of urban sprawl and land-use patterns that leave no other practical option for getting around. Complaining about traffic is complaining about sprawl. One more lane or cars that drive themselves have not been shown to change this relationship.
Sprawl also contributes to greater loneliness and declines in happiness. In 2000, Robert Putnam argued in his book, Bowling Alone: The Collapse and Revival of American Community, that, roughly speaking, every 10 minutes of additional travel time leads to a 10% reduction in social connections. We spend less time with our families, friends, and communities.
There's little doubt that self-driving cars will make commutes more tolerable. But perhaps that's not ambitious enough.
Sprawl is how much of the US provides new housing, and so it's interesting to ask the opposite question: Which cities are actually building new housing in walkable neighborhoods? Here is a study published this week by the Terner Center for Housing Innovation at UC Berkeley that looked at exactly this. What they did was divide all US neighborhoods into five categories based on vehicle miles traveled (VMT) per resident in 2023.
The categories:
Very Low VMT - 12 miles per person per day
Low VMT - 17.3 miles per person per day
Mid VMT - 21 miles per person per day
High VMT - 25.5 miles per person per day
Very High VMT - 37.5 miles per person day
These seem like oddly specific distances, but it's what they used to sort new housing supply. Here's all of the US:
Urban sprawl is how much of the US provides new housing. And here's Conor Dougherty in the New York Times arguing that America needs more of it to fix its housing shortage:
Even if all the regulatory restraints were removed tomorrow, developers couldn’t find enough land to satisfy America’s housing needs inside established areas. Consequently, much of the nation’s housing growth has moved to states in the South and Southwest, where a surplus of open land and willingness to sprawl has turned the Sun Belt into a kind of national sponge that sops up housing demand from higher-cost cities. The largest metro areas there have about 20 percent of the nation’s population, but over the past five years they have built 42 percent of the nation’s new single-family homes, according to a recent report by Cullum Clark, an economist at the George W. Bush Institute, a research center in Dallas.
The obvious benefit is that the resulting housing tends to be cheap. The above article is filled with examples of people buying large homes for a few hundred thousand dollars in newly formed communities across Texas. And if you live in a high-cost city, the social algorithms have almost certainly found you at some point with a shockingly cheap house in one of these places. But, Dougherty also admits that sometimes this may be the only redeeming quality:
Escobar told me he moved to Princeton because he could find a big house there for less than $300,000, but now the city is home, and he didn’t like where it was headed. Over the next four years, he said, his goal is to redevelop the downtown, try to attract offices where locals can work and build out a park system that voters recently funded with a bond measure. “You ask anybody what they love about Princeton, and it’s simply just the affordability,” Escobar told me. “We need to be more than that.”
According to the article, this isn't necessarily a problem, because it's just how cities are built in this day and age. What you do is start with low-cost housing in fringe locations. You grow as quickly as possible until traffic becomes "godawful" and vital infrastructure can't keep up. Then you implement moratoriums on new housing, and start working on other uses like, you know, employment. Eventually, after all this chaos is complete, you end up with something that possibly resembles a real city.
Since the 1950s, new home production in very low VMT neighborhoods has generally been declining. Most of the lower VMT stuff was built before the 1940s, which is why New York City is so walkable and its chart looks like this:
Most newer cities do not build in this way. In fact, based on this study, there are only five large metro areas in the US that have (1) built at least 15% of their total housing since 2000 (meaning, they're a younger city) and (2) built at least 40% of their homes over the last decade in lower-VMT neighborhoods (very low and low).
These metro regions are:
This is not that many cities. At the same time, is it even the right benchmark to be aspiring to? "Lower VMT" just means you don't need to drive as much as you might in other neighborhoods. But it doesn't necessarily mean that you live in an amenity-rich and walkable community. What about the new homes being built in neighborhoods where people don't need a car at all? How many of these exist?
Very few, I'm sure.
Cover photo by Jo Heubeck & Domi Pfenninger on Unsplash
Cover photo by Leon Hitchens on Unsplash
Since the 1950s, new home production in very low VMT neighborhoods has generally been declining. Most of the lower VMT stuff was built before the 1940s, which is why New York City is so walkable and its chart looks like this:
Most newer cities do not build in this way. In fact, based on this study, there are only five large metro areas in the US that have (1) built at least 15% of their total housing since 2000 (meaning, they're a younger city) and (2) built at least 40% of their homes over the last decade in lower-VMT neighborhoods (very low and low).
These metro regions are:
This is not that many cities. At the same time, is it even the right benchmark to be aspiring to? "Lower VMT" just means you don't need to drive as much as you might in other neighborhoods. But it doesn't necessarily mean that you live in an amenity-rich and walkable community. What about the new homes being built in neighborhoods where people don't need a car at all? How many of these exist?
Very few, I'm sure.
Cover photo by Jo Heubeck & Domi Pfenninger on Unsplash
Cover photo by Leon Hitchens on Unsplash
Share Dialog
Share Dialog
Share Dialog
Share Dialog
Share Dialog
Share Dialog
