Why businesses cluster and what's wrong with it

When discussing the potential of Flatbush Avenue to become Brooklyn's destination for the arts, I made a offhand comment slagging business clustering but never bothered to explain myself. Which brings me to the point of this post.

By "clustering" I'm referring to the tendency of businesses to open up near the competition, ultimately creating the Diamond District; stretches of 6th Street and Park Ave. as go-to spots for Indian food; 8th Street for shoes; MacDonald Ave. for kitchen and bath supplies; etc. Chain stores do this as well: Home Depot and Lowe's are practically next door in the Gowanus, as they are in other parts of the country.

The problem with clustering is that it makes people have to travel farther they they'd otherwise need to patronize that type of business. To borrow an example from A Pattern Language: imagine a strip of beach that has an ice cream on the north end. If someone else wants to open an ice cream shop on the beach, they could insure do one of two things: open at the opposite end of the beach, which would likely split the customer base in two by drawing in people at the south end. Or open right next door to the other shop. Either option has the potential of splitting the customer base in half, but the first benefits customers at the south end by requiring less travel.

The classic architecture book refers to this as Hotelling's Law and Wikipedia explains it in greater detail—and why businesses usually choose to cluster.

My point is simply that, given the choice of having, say, Indian restaurants spread relatively evenly throughout the city or concentrated in a single block, I'd vote for the former. Sure, with some businesses—like tile stores—it can be handy to have plenty of choice in one locale. When you care enough, you've always got the option to travel. But often you're not looking for destination pizza; you just want to grab something on the way home from work.

Comments

[jessica]

One positive thing that clustering can do is bring people to an area that they wouldn't have normally gone to before. If Lefferts Gardens becomes known as a destination for "X" (X being something like an arts hub or whatever), we begin to see an influx of people who will come for X and perhaps stay for dinner at a local restaurant, or buy soda at a local deli.

The real trick is in striking a balance between local services (the pizza place, the market, the dry cleaner and Indian restaurant) and the "destination" services.

[danielk]

What you're talking about is not Hotelling's Law, but Economies of Agglomeration. The wikipedia page is weak, but the subject is also discussed in the "Industrial Cluster" page. [[http://en.wikipedia.org/wiki/Industrial_Cluster]]

The study of this phenomena began in the 19th century. Economists didn't understand why businesses would group together when they made the argument you have about reaching more customers and being able to charge local customers more because you saved them the trip to the "Hat District."

To continue the hat maker example, it turns out that all hat makers need similar specialized services. They need access to lawyers, accountants and designers specializing in hats and they need a large selection of materials to go into their hats. By clustering with similar businesses they obtain ready access to this labor. If they were located elsewhere they would need to either be large enough to have these specialists on staff, or they would need to pay more for the generalists available locally to learn their particular business. So instead of being able to hire an lawyer for 5 hours of work, they would need to hire one for 20 hours, 15 of which was simply training.

This is in fact the major drivers of urbanization in modern times. Without this effect NYC would have a population of maybe 15,000 and we would need to wear respirators to go outside because the air everywhere would be too polluted to breathe due to spreading the population thinly over large areas of the planet.

This doesn't mean I don't want services in Lefferts Gardens, nor does it mean that it is a good idea for the government to try to force clustering. They should just try to avoid destroying it with zoning and other coercive measures where it occurs naturally. Silicon Valley might well have been located in lower Manhattan, bringing hundreds of billions of dollars into our economy annually, if our state government hadn't been trying to save the downtown financial district by building the World Trade Center complex and evicting everyone from radio row.

If the apartment buildings in our neighborhood continue to improve and new ones are built the new population alone will support more local services. Restaurants for instance do not benefit from clustering but simply need a mix of locals and foot traffic to support them. This area once had much more shopping back when Rogers still had it's trolley and Ebbets field brought people to the neighborhood in summer; people in their 80's tell me they came here for the shopping. The walk is quite pleasant through the historic district and Nostrand has wide sidewalks that could support sidewalk cafe's and the like, all we need is more people.

Oh, and back to our examples and Hotelling's Law. Imagine that you want a strawberry cone and your husband wants a chocolate cone, but disobeying both Hotellings Law and Economies of Agglomeration the ice cream shop on the North end serves only strawberry and the one on the South end serves only vanilla; now you have to walk to both. If they had obeyed Hotelling's law they would have both had the same flavors, and if they had Agglomerated they would have had different flavors, but both would be at the center of the beach.

[carrie]

Excellent comments! I'll confess that my knowledge of city planning is pretty superficial; I only recently just started reading about this stuff.

I'm currently making my way through Jane Jacobs' _The Death and Life of Cities_ and happened to come across a passage today that seems germane here.

Writing in 1961, Jacobs discusses the then-dismal state of Manhattan's downtown financial district:

This is an immense number of users for a territory sufficiently compact so that any part of it is readily accessible on foot from almost any other part. [The] users represent a tremendous daily demand for meals and other goods... Yet the district is miserable at providing services and amenities proportionate to the need. Its eating places and clothing shops are pitifully inadequate in number and variety for the demands of them. This district used to have one of the best hardware stores in New York, but it could no longer make ends meet and closed.... the district's cultural opportunities are nil."

Jacobs also discusses an early plan of NYC's to destroy Carnegie Hall and move it to a sort of mega-cultural area Lincoln Center--and how that would have been a disaster. (Similarly, she attributes Boston's cultural decline to its segregating out a "cultural district.")

Whenever you have too much of one thing in an area, she's arguing, you end up with lifeless streets. If a neighborhood is nothing but office buildings, you get an area with restaurants that can only survive by catering to lunch crowds; an area that's vacant and scary at night; that can't sustain movie theaters and other forms of leisure.

That said, I don't think we're arguing here. As Jessica pointed out, you can have clusters of businesses AND a thriving area as long as, in general, the neighborhood has mixed uses: not only offices, but residential housing, cultural attractions, etc. We've got a lot of factors here in PLG that make our streets healthy and dynamic, which is probably why people who have been here for a while get irritated by us newcomers complaining about, say, lack of restaurants.