Remember my response yesterday to Randal O’Toole’s Cato article on parking, when I said that I could easily write a three-part series? Not a joke! (Though I might spare you and leave the trilogy unfinished. Maybe.)
Today, I’d like to take on O’Toole’s comments on California, which he argues is too dense and hostile to automobiles to say anything about the real America:
While New York City is very dense, its suburbs are not, so it is not the densest, or even the second or third densest, urban area in America. Instead, that title goes to Los Angeles, followed by San Francisco-Oakland and San Jose—the locations of most of Dr. Shoup’s other examples. Thanks to urban-growth boundaries that are now mandatory for California cities, whatever happens there is hardly representative of much of the rest of America.
He also said something similar in a comment he left on a Market Urbanism post last August about an empirical paper that found that a large portion of the parking in Los Angeles County (population: 10 million) was built because of minimum parking regulations:
I’ve said it before, but Los Angeles is hardly typical of the rest of the U.S. It is the densest urban area in the country (and not just the city is dense). Beyond that, my more important point is that developers build parking lots everywhere, not just where there are parking minimums.
Average density: a misleading measure of walkability
My problem here is that O’Toole is using the literal definition of “density” – that is, average density. But this is just a shorthand for what really matters when you decide whether you need a car or not (and developers decide how much parking they need to build to maximize profits): walkability and access to mass transit. We often use “density” as shorthand for auto-orientedness, but it only really serves as a good metric over very small spaces. When you start looking at metro areas, though, its utility declines.
Paul Mees and Jarrett Walker have written about this to death, and it can get a little tricky to think about, but I think the best way to grasp it is with an example. Say you have two metro areas with the same population and same area, but with one where everyone’s concentrated in one city with only few people scattered around the suburbs, and another where everyone’s the same distance from each other. They’ll both have the same average density (population divided by area), but clearly one will be walkable and one will not be. Obviously this is a stylized example, but similar dynamics inevitably play out in the real world. California’s suburbs may be dense, but they’re still built in a very suburban style and are thus largely unwalkable. Much of this effect is achieved through separation of uses and the road network: Even if you live at Manhattan densities, you’re going to need 100% parking if the road network is all cul-de-sacs and limited-access highways with low connectivity and mixed uses are not allowed.
All of this is to say, Los Angeles’ high average density seems like a flimsy reason to disregard one of the few (two, by my count – the other looks at Queens) empirical studies on the effects of parking minimums. And while it’s true that California has urban growth boundaries that make it relatively unusual among American cities, it’s also true that the same anti-growth environmental forces also put in place some pretty anti-growth policies in the already built-up areas, so it’s not at all clear that the net effect is to make the place less car-oriented. I’ve never spent any appreciable time in the state, so maybe I’ve just been deceived by Hollywood and everyone I know and everything I’ve ever read, but I’d bet that it is indeed at least of average auto-orientedness for American metropolitan areas. If density in Los Angeles County is hobbled by parking minimums, then I see no reason to think that the same wouldn’t apply for similar parking minimums in metro areas throughout the rest of America.
Where do you get that (by any rational standard) anyone thinks that Southern California suburbs (for god’s sake it’s almost all suburb) are dense? What an absurd idea. O’Toole is hardly anyone to judge urbanism.
In LA there are a few small nodes of walkable urbanism but most of it — in fact all of California, in fact the whole west coast with very few exceptions — is suburban i.e. not walkable.
Density per se is irrelevant. On both sides of the debate. Too many Urbanists salivate for “density” and what you get is discussions with folks like O’Toole i.e. a diversion od energy.
Southern California is the second densest metro area in the US, counted properly (i.e. weighted, not average). It’s not what you’d expect, but most of LA-ex-mountains, the Valley floor, and some of the inner suburbs are fairly dense, so it overall adds up.
The thing is, density doesn’t have to be pedestrian-friendly. There’s some literature on how a good walkable neighborhood needs to have some density, but a dense neighborhood need not be walkable or transit-oriented; one paper I may be able to hunt down gives West LA as a specific example of pedestrian-hostile density. There are even more egregious examples: a post on Human Transit looks at fairly dense condo areas in Vegas, whose metro area density is again higher than you’d expect, whose street access passes through just one or two chokepoints, making walking too cumbersome.
Yes, I think we are agreeing; maybe I didn’t put it correctly.
The larger story is that in terms of politics we pay to much attention to “density” and should spend our time creating interesting, comfortable places:
“Density if a Byproduct of creating interesting places.”
Read, for example:
http://citycomfortsblog.typepad.com/cities/2005/05/density_again.html
http://citycomfortsblog.typepad.com/cities/2009/12/oh-god-not-again-the-density-debate.html
http://citycomfortsblog.typepad.com/cities/2011/02/the-war-on-density-publicola-seattles-news-elixir.html
I suppose if you take it as a given that 1) More density is better and 2) Everyone must have a car and drive everywhere, you end up with the situation where traffic is a huge problem, and this leads you to devote as much area as possible to moving cars, to the detriment of everything else. Result: dense and unwalkable.
“Say you have two metro areas with the same population and same area, but with one where everyone’s concentrated in one city with only few people scattered around the suburbs, and another where everyone’s the same distance from each other. They’ll both have the same average density (population divided by area), but clearly one will be walkable and one will not be.”
Sorry for plugging myself, but you don’t have to reinvent the wheel here:
http://www.austincontrarian.com/austincontrarian/2008/05/another-feature.html
http://www.austincontrarian.com/austincontrarian/2008/09/the-association-between-density-and-mode-of-commuting.html
http://www.austincontrarian.com/austincontrarian/2008/09/barnes-weighted-density-stats.html
Alon’s right, though, that dense places need not be walkable. A neighborhood made up of Le Corbusier towers divided by highways can be quite dense but deadly to walk.
If you follow the link to the Human Transit post, you’ll see that you’re cited indirectly in the post, and directly in the comments – as is Gary Barnes.
yeah, thanks. I think any discussion of weighted density and transit use should start with Barnes, who calculated weighted density specifically using traffic zones (and whose analysis was a lot more sophisticated than mine).
O’Toole claims that New York’s suburbs are not dense? The state of New Jersey is more densely populated than Japan or India!
(See how easy it is to make valid-sounding statements of spurious actual value, when you use too coarse a geographic area over which to compute density?)
Your point about the imperfect correlation between density and walkability is an important one. We at New Jersey Future have long joked about “dense sprawl” (of which L.A. is perhaps the prime national example). When I first joined the organization 11 years ago, I was bothered enough by the constant citation of New Jersey as the most densely populated state in the nation (which is true, but essentially meaningless at the scale of individual developments or even cities) that I eventually wrote a 30-some-page paper about the right and wrong ways to quantify density. You might be interested in looking at it — the intro page is at http://www.njfuture.org/index.cfm?fuseaction=user.item&ThisItem=53&ContentCat=3&ContentSubCat1=14&ContentSubCat2=3.
Of particular interest is the section on “Density and Traffic”, starting on p.16 (labeled as p.14) of the pdf, in which I even give credit to L.A. for being the most densely populated urbanized area in the country. (I concluded the section by saying that “Los Angeles should stand as a stern warning against developing simultaneously at high densities and in an automobile-dependent style.”
Just as O’Toole uses statistics in other scenarios to distort true meanings, he does here also. The numerical density of a broader metro area is less relevant than the density–and really the built layout–of its main commercial areas. And it doesn’t control for factors like open space and land boundaries. New York may be less dense than LA overall, but all of Manhattan, and perhaps even Brooklyn, are denser than probably any one LA neighborhood. Thus you can’t compare the two when talking about what each region needs to do for transportation…