29 Comments

Orthodox Jews have a solution. Since they go to synagogue on the Sabbath and cannot use any machine to get there, they cluster around synagogues. A side effect is the sort of get-togethers and mutual support that you wish for.

Expand full comment

Aha – indeed. In fact, I have written a paper on just this... https://www.createstreets.com/wp-content/uploads/2021/12/Tottenham-Paper-1.9.pdf

Expand full comment

This is definitely something I'm thinking about as I move right next to my brother/sister-in-law but not right next to my friends. I wish we would all just live right next to each other! C'mon people! But the externality point makes a lot of sense. Maybe it's even like a common pool resource. My moving away takes away from your resources.

Expand full comment

This is an interesting question to ponder. I moved last year to be closer to friends and my quality of life has definitely gone up-- the proximity was ultimately more valuable than the other amenities of the neighborhood I was in originally.

What nudged me towards the decision was taking a closer look at how I spent my time, and realizing some of my best days were when I was in the neighborhood where more of my friends are. Maybe if there was more focus culturally on spending time optimally rather than the consumption of other amenities, more clustering would happen?

Expand full comment

What I see works in the real world is to create high-trust subcultures with high barriers to entry, and then to exist as a member of a subculture rather than as a member of a friendship group. For example, gay men in the 80s in NYC would probably not miss clustering with their hometown friends if they were part of the subculture. This has the added benefit of essentially being portable, as the traits that get you entry into subcultures work anywhere, and benefitting from scale as barriers can be higher in bigger populations.

So maybe the solution is to be an exceptional person in a subculture, and normal people lose.

Expand full comment

I've had a very similar thought; as we get older and seek to expand and deepen our relationships, maybe it's best to meet people through a shared interest.

Expand full comment

Yes, why do people think so little about friends when they choose to move? Why do they presume that they can find them trivially but they won't be able to find other things and need to plan around them? Or why don't people make new friends near to them and choose to stay near them? Why is it so easy to move for work, school, but not friends? Or why don't people make friends with people near them in a stable way, and instead find 'loving their neighbor' rather near impossible?

Expand full comment

When I lived in New York I lived in Downtown Brooklyn and so did many of my friends. So there are some exceptions! Of course, Downtown Brooklyn is a bit of an anomaly. A host of new buildings have sprung up in the last 10 years thanks to tax breaks/rezoning, making it a relative bargain to live there if you work in downtown Manhattan. And it helped that my friends were a relatively homogenous bunch -- we were all young professionals with no kids.

Expand full comment

I agree wholeheartedly with your essay and think about this problem a lot!

Perhaps you've read this but Eliezer Yudkowsky covers similar territory here: https://www.lesswrong.com/posts/cmiRk9XtT9Psnd3Yr/movable-housing-for-scalable-cities

particularly benefit #3

Expand full comment

The way I solved this in San Francisco was to advocate for one area as the best square mile: https://bit.ly/neighborhood-location, https://goo.gl/maps/2Mpaw2213uwkrx3q7

Then all the people that were open to living in the Bay anyway moved there, kicking off a little movement that caused 150ish people to move to 2022-2023 and incentivized community startups like the Commons (which I helped start), HF0, Solaris, and MadSci makerspace to locate there, which created a positive feedback loop.

Another way I’ve seen it solved is for a critical mass of core people in a group share a duplex or a triplex. Once you’ve got a nucleus of 2-4 key people together, then more peripheral people in a group can move closer too.

Expand full comment

Successful cities pull friends apart. They match people with their perfect job. But each individual's perfect job will be distributed somewhat randomly around the city. If friends do co-locate for a while, job matching will pull them apart eventually. Only solution as you say is to make new friends wherever your job lands you.

Expand full comment

I don't agree with this. You almost never have to move location because your job has moved. Jobs are usually central (or central-ish at worst) so you can flex 5-20 mins on your commute. I have lived in 7 or 8 places in London now, including my parents' house deep in the suburbs, and I have never moved for a job – in fact my job has only rarely even constrained my choice, and usually just to zones 1-2. Only people with strange job locations (e.g. doctors) would usually have this problem. This is the point of a city! It's usually considered a bonus if you live under 30 mins commute to your job, rather than a requirement. The requirement is living in the same city.

While this job issue could theoretically have constrained me if all of my friends were doctors or similar, in fact they have all always worked in zone 1 central London for the last ten years so it never would have been a problem in practice.

The two things are compatible because of the number of trips you need to take. Commuting to work is once there, once back per day. Whereas the lower the better for friends, as you may take a great many trips. The issue is, as I argue in the post, the spillover benefits of any given friend moving near the cluster are not captured by that friend.

Expand full comment

London's jobs are pretty centralised as cities go. Contrast to car-dependent cities where jobs are dispersed much more widely. To take LA for example, it's not practical to commute to most jobs. Angelinos tend to live on the same side of town as their work. Would your position on London change if there were lots of jobs in Richmond or Woolwich?

Also the bigger a city gets, the better it gets at matching, the worse this problem gets.

Your point about spillovers and externalities stands. But I think it's always in tension with matching in job markets. Especially when a city gets really big.

Expand full comment

LA is atypical and if they sorted out land use regulation I expect it would revert towards normality. London is like most cities in the world in this respect. There would never be loads of jobs in Richmond and Woolwich unless we did something stupid to force them out there. But yes, if jobs were distributed widely, then there would be another factor pulling groups apart. Still, the solution would just be to live in the middle. If the jobs are spread around, living in Farringdon or the City would keep your cluster near most of them.

Expand full comment

But in car dependent cities that's exactly what happens –  jobs naturally disperse around the city. It's a function of the transport system IMO, not planning. You don't have to go to Atlanta or Houston to see it. In Dublin there are a lot of good jobs in Sandyford and also City West. From where I am, pretty central on the Northside, they're 1hr 15mins away in the morning!

Expand full comment

Even in car dependent cities jobs disperse much less than people. But I accept that to the greater the extent that jobs are (for some reason) distributed around the city, to that extent we will also see job concerns pulling friend clusters apart. I dispute that this problem is very substantial in most major cities, but there are no doubt some where it’s important.

Expand full comment

Also living in the middle – definitionally not a solution for everyone

Expand full comment

Not even true. You imagined a city where the jobs are evenly distributed around. That would create a demand for housing in the centre. Put enough of it there and we could all live there if we liked. (Cities like Manchester and Leeds had zero residents in the central square mile during the 1980s and 1990s and the City has only a few thousand today. It would be just as possible to do the inverse.) But an inside-out city is atypical because the jobs usually want to go in the middle because job agglomeration scales more than home agglomeration.

Expand full comment

In my experience the problem has been a combination of the quirks of when people move, the availability of apartments and especially willingness to pay (in the context on NYC) that has kept me and my friends from coordinating better.

The thing about the 'be in the middle' theory is it's flat out wrong, as you observe - the utility of proximity is roughly an inverse square, so if there's two useful things A and B, the midpoint is actually the worst possible place to be, you want to be at A or B instead. So it takes a weird config combination of A+B+C to make picking such a point wrong.

Expand full comment

Interesting. Well, it would be amusing if my theory was completely wrong because of the point you make, and it would be delightful if it was as simple as making housing much more abundant.

Expand full comment

There's also the fact that your friends have other friends outside of your social group. Even if you could persuade your friends to consider moving closer to you, unless you're highly unusual or in a cult, they are going to have to consider proximity to many people you do not know or are only vaguely acquainted with.

Expand full comment

I wonder if this is true more broadly. When I and partner lived in East London (contrarians, I guess?) most of the Aus ex-pats I knew lived in Hammersmith or Shepherd's Bush. Not identical to what yr discussing, but this being up to your elbows in countrymen halfway around the world is the opposite dilemma to the one you describe. Anyways, just a thought.

Expand full comment

I think expat communities may be a rare type of community that do successfully cluster. They are relatively small, much more different to the average person, and have particular cultural amenity needs.

Expand full comment

I completely agree, which is why a group of friends in San Francisco have been clustering into a single square mile, along with 200+ other people: https://jasonbenn.notion.site/6-Becoming-Hyper-Local-Real-Estate-Guy-0e011a05e02a4df79e2355ed0789123b

I think there are two insights that make this project possible:

1. If you choose a location where people already demonstrably want to live (the single best square mile in a tier 1 city), then recruiting becomes easier

2. If you expand your ambitions to 200+ people instead of just 10 best friends, it becomes paradoxically easier to coordinate. Friends of friends feel more welcomed, and when it's literally remarkable then people will talk about it.

I'm happy to walk you through how we've gotten here in the last 6 months, I'd love to see something similar happen in the rest of the world's great cities. jasoncbenn@gmail.com.

And shoutout to Andy Jones for pointing me here in his newsletter! https://andyljones.substack.com/

Expand full comment

Excellent. I will keep my eyes peeled for more information on this, as it's a great initiative. Both your (1) and (2) seem like sensible points.

Expand full comment

Any "when": Something perhaps more common a hundred years ago, for cultural icons, aristocratic or upper classes - more socializing may have been organized around country houses, or areas such as the Hamptons, Martha's Vineyard, etc. These areas provide venues to solve the problem. There were more downscale places all up and down the Eastern Seaboard, as well, where families would see each other year after year. It's the workaholism required even of so many professionals and others trying to achieve or retain high status that makes workaday geography so problematic for a human life.

Expand full comment

It does seem that classes can do it, but individual friendship groups seem to find it harder (today at least).

Expand full comment

Non-native-language-speaking immigrant communities are a counterexample - many cities have a Chinatown. I suspect what is going on is that economic quantifiables (e.g. job prospects, good schools) are overweighted in location choices whereas intangible final consumption goods (e.g. proximity to friends, parks) are underweighted.

Expand full comment

Agreed.

Expand full comment