« Seawall, See Wail | Main | Medical Cost of Obesity »

August 15, 2005

Sprawling Alone

In Bowling Alone, Robert Putnam argues that the decline of social capital in the United States can be attributed partly to urban form. In other words, according to Putnam, sprawl is at least partly to blame for the present derth of bowling leagues. But is it really?

Putnam's arguments (summarized at the end of Chapter 12) are threefold.

  1. "Sprawl takes time" and results in more time spent alone in a car and therefore less for civic engagement.
  2. "Sprawl is associated with increasing social segregation" and that segregation has led to less community participation.
  3. "Sprawl disrupts community 'boundedness'" and that physical fragmentation reduces societal involvement.

Although Putnam's claim--that sprawl erodes social capital--is widely referenced, my survey of the evidence makes me suspicious. My objections are threefold.

First, sprawl does not absorb more commuting time than does urbanization. Data from the National Household Transportation Survey shows astonishingly similar travel times across residential densities. Actually, residents of high densities spend the most time traveling. (Caveat: the NHTS is for "travel time" not specifically "commuting time," which Putnam is interested in, though commuting only accounts for about 1/4 of all personal trips.)

In fairness, high density households spend about 1/3 less time driving and proportionately more time walking, busing, or biking. So it's possible that urbanites use that extra 20 minutes per day to form social networks on public transit, but it seems equally possible that suburbanites form social networks while carpooling. In any case, no matter what the residential density, households sink roughly the same large chunk of time into commuting (74 to 79 minutes/day, most of which, even at high densities, is driving).

Putnam asserts a rough formula for measuring the effects of commuting: "each additional 10 minutes in daily commuting time cuts involvement in community affairs by 10 percent." Still, I can't see why sprawl is the culprit here. Instead, the culprit seems to be something like congestion, or perhaps the sheer physical size of cities (admittedly, related to sprawl), or perhaps even the population size of cities, which necessitates physical breadth. If sprawl is irrelevant here, it may help explain a point that Putnam apparently takes to be puzzling: that both surban and urban residents in big metro regions have less social capital than their small town counterparts.

Second, sprawl may be associated with social segregation, but the evidence that this erodes social capital is not conclusive, at least as far as I'm aware. Putnam does cite a couple of interesting studies here, but there are many more he bypasses. In a more comprehensive survey of the evidence, "The Effects of Sprawl on Neighborhood Social Ties," Lance Freeman finds that "the existing evidence is not conclusive" and that moreover very high densities may actually be corrosive of social capital.

Freeman's study goes on to find that residential density is unrelated to neighborhood social ties, but is strongly related to automobile dependence. As car dependence is generally a feature of sprawl, it may be that Freeman's study supports Putnam's conclusions. Still, both Freeman's survey of the literature and his data analysis should serve as a cautionary tale for imputing too much explanatory power to low residential density, which is often treated as the defining characteristic of sprawl.

Third, the importance of community "boundedness" is, as far as I can tell, based on only one study that's now more than 30 years old. Admittedly, it was something of a classic, but it's rather hard to imagine that the same cultural and geographic forces in play in 1972 are the same ones that now impede social capital. For just one example, city center populations have been growing again, rather than hollowing out as they were in the 1970s.

I'd like to see more evidence on this subject. It could be that Putnam is basically right and I'm just nitpicking, but for the time being I'm suspending judgment on the social effects of sprawl.

Posted by Eric de Place | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Sprawling Alone:


Eric: I am also skeptical of Putnam's claims. I don't doubt his statistics which apparently show that America is becoming less social. But I doubt the extent to which it is a factor of geography.

At this moment my wife and I are living in Central Texas while she finishes her residency and then we plan to return to the Portland area. So I've been doing a lot of thinking about what kind of community we want to live in and I've also been doing a lot of thinking about where we live now. We live in what is perhaps a classic "exurban" subdivision northwest of Waco near a town called China Spring TX. It's the sort of place that David Brooks of the NYT seems to be enamored with. Wide streets, big houses, big lots, lots of SUVs and Republicans. Most commute the 15-20 minute drive into Waco and a few hardy souls make the 1.5 hour drive to Fort Worth on a daily basis. There are also lots of kids and stay-home moms.

And you know what? It is not only the most social neighborhood I've ever lived in, it is one of the most diverse. And I've lived in Eugene, Portland, Seattle, Juneau, and Washington DC before moving here. The five immediate neighbors that abut my property are an African American family, a retired French couple who don't speak English, an Argentian businessman, an Italian-American family who just moved here from Memphis, and a Mexican-American family. They are always organizing neighborhood parties. During the summer my kids roam freely from house to house and all the neighbors seem to keep an eye on everyone's kids. Social events revolve around the high school sports and youth soccer. In the short time we've lived here I've gotten to know far more of my neighbors than any place else I've ever lived. Down the street there is a family in which the husband is currently deployed to Afganistan and the wife recently gave birth to a very premature baby that takes intensive 24-hour a day care. The other moms in the neighborhood just automatically took in their 3-year old daughter and took care of her during the long weeks that the wife was with her preemie in the hospital and continue to baby sit for her every day and do all her shopping while she stays at home taking care of the baby. They weren't asked to help. Everyone just automatically pitched in. That sort of thing NEVER would have happened in the SE Portland or NE Seattle neighborhoods I used to live in when single.

OK, this place would bore anyone single to tears and I would have probably rather put a bullet in my head than live here when I was single. But that's not the point. When I talk to people about whey they moved out here one of the things I hear the most is that they did it for the sense of community. People are fleeing the larger urban areas in search of community. If you look at an area like Dallas-Fort Worth you will see that it is surrounded by literally hundreds of smaller independent towns where people are moving to in droves, in part, because they are searching for a more manageable sized community where they can feel like they have some "ownership" in the local schools and other aspects of the community.

Now there are certainly a lot of negative aspects to unplanned sprawl. The environmental effects such as habitat loss, the increased energy use, increased congestion etc. etc. There are plenty of reasons to oppose unregulated sprawl. But I think it's stretching the argument to blame sprawl for the breakdown of social fabric in this society.

Posted by: Kent | Aug 16, 2005 7:36:54 AM

Hmmm, interesting. I recently moved into a SE Portland neighborhood, and while it is definitely not the most diverse place, it is very neighborly, in the very way that you describe. If I had to take a stab, I'd say it had to do with the high percentage of owners vs. renters.

Posted by: colorless green ideas | Aug 16, 2005 11:11:09 AM

Every house I lived in in Portland was a rental. Mostly within 2 miles of Reed. Now that I think about it, I wonder if there is a different level of neighborliness that occurs between owners vs renters. In other words, I would assume that the "permanent" homeowners in each neighorhood know who each other are and probably tend to socialize with each other much more than they might with the renters in the neighborhood. Owners do have a lot more to talk about than renters because they have a lot more invested in the neighborhood.

The homeowner's association in my own particular neighborhood bans rentals so everyone here is an owner. If you see someone in a house you know they are the owner. That certainly keeps out the speculators and investors who want to flip properties, although that really doesn't happen here anyway as land and construction costs in most of Texas are too cheap to support a housing bubble.

Posted by: Kent | Aug 16, 2005 12:06:04 PM

commuting only accounts for about 1/4 of all personal trips.

Do you have a citation for that? How is "personal trips" defined -- any trip that's not business-related? The travel surveys I've seen (for instance, for the NY metro area) indicate that commuting accounts for 10% of all trips and 15% of all miles traveled.

Urban and sprawl neighborhoods can be equally neighborly, but there are certain forms of community interaction that are more difficult and sometimes impossible in sprawl. They include spontaneous meetings and interactions on foot, on porches and other "intermediate" spaces. Gatherings and meeting of new people in village greens, town squares and city plazas. Street musicians and other informal performances. Street vendors and informal markets.

Posted by: Laurence Aurbach | Aug 17, 2005 8:58:38 AM

Commuting time depends upon the spatial structure of the region, so averaging commute times nationally is problematic [expanding on Eric's point]. This gets back to the density of employment centers post - dense, polycentric employment centers may reduce TT better than a monocentric employment center [which makes this post - http://cascadiascorecard.typepad.com/blog/2004/10/sprawl_rebounds.html - interesting, as we know Seattle is polycentered].

Income can influence commuting as well - affluence means you can work at home sometimes [e.g. http://www.psrc.org/datapubs/pubs/trends/t21nov03.pdf
pg 3 fig 1].

And women find that commuting and running errands is an issue, as the services they utilize are often scattered, requiring extra trips [this supports the 'sprawl is bad' thesis].

I'm trying to make Buckley a TND town, but I know that King/Pierce Cos are polycentered so I'm still going to have long commutes but hopefully the folks here can squeeze in a few walks per week...hmmm...dog facilities to promote walking...hmmm...

Posted by: Dan Staley | Aug 17, 2005 9:31:30 AM

Laurence-- that figure comes straight from "Bowling Alone." It's on page 212, in chapter 12. He actually says, "Commuting accounts for little more than one-quarter of all personal trips..." I've heard various estimates for commuting as a share of all trips, some higher and some lower. I'm not sure what definition Putnam is using and I suspect it's defining the term that's behind the various estimates.

Also, I agree that certain forms of interaction are more difficult in a sprawling setting, but it appears likely that some feature of sprawl is offsetting those difficulties(or that, conversely, some aspect of the urban form is inimical to interaction) because the studies I'm aware of are pretty darn inconclusive about the impact of suburb/city on social interaction. Which leads me to...

Kent-- Thanks for sharing your story. Your experience is anecdotal evidence that good community building can happen without urban-style development. I'm actually now quite curious about the role of "boundedness" in developing social capital.

Could it be that your community is discrete/seperate enough that people bond together more than if there were an undifferentiated sea of sprawl? Or that a city neighbhorhood with a distinct sense of cultural or geographic identity tends to bond more closely? The study Putnam uses is, as I pointed out, more than 30 years old. I wonder if anything new has been done on this topic.

Posted by: Eric de Place | Aug 17, 2005 10:08:32 AM


Here in this part of Texas there is pretty much endless space for housing development so it tends to get funneled first into the choice school districts and second, around interesting geography. The Waco metro area is about 250,000 so it's about like Eugene or Salem. But they have never consolidated the school districts here so there are perhaps 15 or more semi-rural and suburban school districts in the region. Even within Waco city limits there are five. Because of all the testing in Texas it's pretty easy to track which are the higher performing schools and the new nicer developments tend to cluster within those school districts. Probably only three of them. And since most of the people moving into these newer suburban and exurban developments are families with kids, the identity with the schools is immediate and strong. Pretty much the entire community shows up for football games. That never happened in Eugene where I grew up and where the entire city was one big school district.

And I also think the outlying communities are discrete enough still. Not one giant blob of development like you see for example, between Seattle and Tacoma. Right here they tend to follow the Brazos River which is very crooked and it's the edge of the hill country so the nice areas have lots of trees and views. This is the subdivision where I live:


It's isolated enough and quiet enough that people are always out walking their dogs and kids every evening so you meet your neighbors and if you are working in your yard they are always walking by. And the folks who run the homeowner's assoc are always planning block parties, progressive dinners that move from house to house, weekend neighborhood cleanups of the common areas, that sort of thing.

Posted by: Kent | Aug 17, 2005 6:57:29 PM

If you read Putnam's book carefully you will see that his chapter on sprawl is actually a look at a POTENTIAL explanatory factor for the decline in social capital. He says that it actually accounts for less than 10% of the decline, ALONG WITH a host of other, related factors with which it is correlated (e.g. time and money pressures on two-career families) (see page 283 for a summary). To understand his rather complex and exhaustive data sources and analysis one must read the Appendix to his book.

Posted by: julia | Aug 20, 2005 10:14:07 AM

Perhaps the issue here is not sprawl itself but the typically short length of time people live in one place these days. According to James M. Jasper, author of Restless America, "in a typical five-year period, only about half the population (53 percent) is living in the same place at the end as at the beginning."


In the two years we've lived in our house in the Mt. Baker neighborhood of Seattle, four out of ten houses on the block have changed occupants and/or owners (five if you count us buying our house), and two have gone vacant as their owners have moved to nursing homes.

A certain amount of that moving is job-related, but I'd guess a good bit comes from having to move because a house no longer accomodates the owners' needs.

Maybe we ought to be thinking about what would make it possible for our housing, for starters, to accomodate aging in place. For example let's say we build (and allow in zoning) family houses with smaller detached accessory dwelling units. While raising the kids, the family rents the ADU to singles--or even, a mother-in-law! (Could work if it was detached....) After the kids move out the couple moves into the ADU, remaining in the community rather than having to relocate in order to find age- and family-size-appropriate places to live. (To high-rise condos downtown, if Clark has his way....;-))

Posted by: Rob Harrison AIA | Aug 29, 2005 1:52:53 PM