A few weeks ago, the National Association of Home Builders (NAHB) released a study titled, Housing Preferences of the Boomer Generation. I learned about it through a reaction piece posted under the clickbait headline: "New Study Confirms Baby Boomers are Clueless." The article is a bit sensationalistic but better than its headline suggests, and makes some valid points. In particular, it takes strong issue with the following finding:
It appears that what boomers want are big suburban houses on winding culs-de-sac.... It’s bizarre. 78 percent actively prefer a cul-de-sac to a connected street. They want double car garages. They want 2,000 square feet on one level. They want three bedrooms. And they really, really don’t like the city.
The author makes the point that such a living environment is antithetical to aging in place, and promotes social and physical isolation among seniors as driving becomes a more onerous task. This is a topic that Strong Towns's own Jason Schaefer explored in depth a few weeks ago (motivated by the same survey I'm writing about now). Rather than rehash it, I'll link to it here.
What I want to do is expand a bit on why this survey and others like it have results that so often appear contradictory.
For one thing, the average layperson doesn't spend their time thinking about urban design and its implications. To those of us who do—the planners, engineers, involved placemakers and urban enthusiasts who make up some of the Strong Towns readership—this is easy to forget. But the cul-de-sac vs. connected street grid question explicitly references "traffic flow"; i.e. it basically sounds like it's asking, "Do you want a lot of traffic in front of your home?" All else is presumed equal. When you look at it that way, there's a "Duh" answer to the question.
What's not valid is to read into this the idea that suburbanite boomers, or members of any other generation, don't want the advantages that come with a more connected street grid: less need for traffic-clogged arterial stroads, greater walkability, and the possibility of retail within easy walking distance. In fact, we know many baby boomers, the generation that embraced the suburbs most wholeheartedly, want retail that's walkable from their homes, because they tell us they do: It's rated right below "typically suburban" among their top priorities in the NAHB study.
Of course, "typically suburban" nearly always means "not walking distance from retail" in the real world, so there's a glaring contradiction staring at us from the top of these findings. But is the issue that boomers are collectively clueless and don't understand this contradiction? Or is it that the survey format doesn't capture the way we actually making housing and location choices?
In general, survey questions in which respondents are not prompted to consider trade-offs—i.e. acknowledge the fact that all else is never actually equal—are often fairly useless. For example, solid majorities of people will answer "Yes" to the question, "Do you want lower taxes?" But solid majorities will also answer "Yes" to, "Do you want more and better government services?" It's only when you frame it as a trade-off between the two that you start getting any real insight into people's political preferences.
There is a lot of incoherence in how we talk about Americans' housing preferences, because there's a lot of incoherence, I suspect, in how Americans think about their own housing preferences. The reality is it's not so simple as asking survey respondents to separate things like "living on a cul-de-sac" from "living in a suburb" from "living in a safe community with good schools" from "living in a car-dependent place" from "living in a convenient place." Not in terms of their own life experiences and the choices available in the cities they've lived in, which have no doubt, required trade-offs.
Economists use hedonic pricing models to study housing desirability. These models attempt to predict the price of a housing unit based on a complex bundle of attributes, each of which has a certain value, or utility, to the buyer. Think of the utility as "How much of a premium would you be willing to pay, all else being equal, for a house with a garage versus no garage? Or for a house near a park versus a house not near a park?"
A hedonic model for housing might treat the price as a function of many variables in three categories: characteristics of the house or housing unit itself, locational advantages or disadvantages, and neighborhood amenities (or disamenities). For a (necessarily incomplete) list:
- Size in square feet
- Number of bedrooms
- Number of bathrooms
- Garage present (yes/no), big enough for how many cars?
- Fireplace present (yes/no)
- Patio present (yes/no)
- Distance to downtown
- Number of jobs within 30 minutes by car
- Number of jobs within 30 minutes by transit or other modes
- Distance to nearest grocery store
- Distance to nearest park
- Quality of local schools
- Crime rate
- Presence or absence of a Homeowners' Assocation or deed restrictions
- Racial / ethnic mix
- Average and range of household incomes
- Age profile of area residents
Imagine it were possible to rank those amenities from the above lists that you most care about, and then simply go shop for a house that delivers the perfect combination of the things that matter to you without having to pay for the things that don't. I value amenities in walking distance very highly, for example. I don't care at all about having a two-car garage, or a particularly large house—1000 square feet for me, my wife, and our dog and two cats would be perfectly cozy. I don't feel strongly about whether I live in a detached house or an apartment. I do care about a low crime rate. I do want to avoid living right on a stroad. And so on.
Here's the catch. You can't get just any old package of amenities.
You know how your cable TV service is bundled? You're required to purchase a package of channels which includes many you will likely never watch. You can't pick and choose a la carte. There's all sorts of niche programming that is only profitable because you have to pay for it if you want ESPN or Comedy Central. How many people who subscribe to the Hallmark Channel (because it's part of the bundle) actually like it?
Housing is similarly "bundled"—you can't disentangle a lot of these attributes from each other. And there are many, many combinations of housing or neighborhood attributes—think of them like channels—that homebuyers might love, but homes fitting the description are few or nonexistent. Sorry, you can't get this with that. Yet the housing market isn't dominated by an oligopoly of a few big companies like television is. There are millions of sellers and buyers in the housing market, and millions of individual lots on which housing exists or could be built, so what gives?
In some cases, the answer is that the "bundling" is inherent to the attributes in question, because they are inherently linked: for example, urban form itself affects the ability of your city to deliver walkable amenities. Culs-de-sac and large lots put a cap on the achievable population density.
It's estimated that a full-service grocery store has a market area of close to 10,000 people—that is, for roughly every 10,000 people, your local economy can sustain one grocery store. (Estimates vary, but here's one study by consulting firm Metropolitan Research and Economics that estimates 8,000 to 9,500.) What's a reasonable walking distance to a grocery store, assuming you have to lug groceries home with you? Let's say half a mile. A half mile radius around a store has an area of 0.785 square miles. To fit 10,000 people in this area would require a population density of 12,739 people per square mile.
For reference, this is slightly more than the average density of Chicago (11,864) and less than only Boston, New York City, and San Francisco, as large U.S. cities go. Thus, only in that handful of cities is it feasible for most residents to have a full-service grocery store within walking distance, based on these estimates.
So yeah, your preferred urban form is inherently at odds with your preferred neighborhood amenities. Unless what you're saying is actually, "I want a house on a cul-de-sac with a large yard on a quiet street, but nearby areas should be dense enough to support walkable retail." That is, "I want all of the benefits of density with none of the drawbacks." That's nice. Me too. I'd also like an all-expenses-paid trip to Italy.
My suspicion is that these survey questions are basically understood by readers as, "Would you like to have your cake and eat it too?" Why, yes! And yet they're cited by both developers and pundits as evidence for cherry-picked conclusions. Americans love suburbia. Americans hate suburbia. Millennials don't want to drive cars. Millennials still love cars. Urban millennials will all pack up and move to the suburbs as soon as they have kids. Or they won't. Be skeptical.
Ask About Trade-Offs
When you build a neighborhood full of those basically private streets like culs-de-sac, they start imposing a significant negative externality on everybody. You lose any hope of walkability. You need wider and wider arterials to deal with all the downstream traffic. You end up with a car-dependent urban form, and a predictable congestion problem.
But these surveys aren't asking that. If you want to know whether people actually like that trade-off, whether they think that trade-off (car-dependence, congestion and sprawl for a tranquil, private immediate environment) is worth it, you need to actually ask them about that trade-off.
How policy and history have increased the "bundling" of housing options
There's another problem. We could have a far more "unbundled" mix of housing and neighborhood choices without the policy distortions that have led to our suburban status quo. For one thing, FHA lending requirements privilege very specific building types and neighborhood types.
Additionally, white flight, racial tension, and macroeconomic forces that eliminated working-class jobs —i.e. all the factors that led to urban decline in the mid-to-late twentieth century—meant that neighborhoods that might be very desirable on the H and L axes of the hedonic pricing model (housing features and locational attributes) became extremely undesirable on the N axis (neighborhood amenities, like good schools and low crime). Many people who could leave made a rational choice to leave, albeit a choice that did grave damage to the neighborhoods they left behind. Many of these neighborhoods have wonderful, historic housing stock to this day, but endemic poverty and blight.
You can and should fault the social implications of this, segregation and concentrated areas of poverty and disinvestment. But these forces, like it or not, do continue to affect many people's perceptions of urban areas, particularly the baby boomer generation. Look at how "urban" is a code word in our politics, most often a code word for "black and poor." Can we really look at a bunch of white boomers saying they don't want to live in an "urban" area and conclude that they're expressing an opinion about urban design, architecture, or walkability? Get real.
A high-school student I tutored once in a wealthy Silicon Valley suburb, when he heard I lived in San Francisco, told me "Eww. Isn't San Francisco kinda ghetto?" Was he reacting to the Victorian "painted lady" houses and the ethnic restaurants and boutiques? I doubt it.
Zoning is another culprit. In most parts of most cities, zoning rules explicitly disallow neighborhood typologies that might represent "intermediate" forms between urban and suburban. All the yellow on this map represents areas in Minneapolis zoned exclusively for single-family homes—not even duplexes in most cases due to unrealistic minimum lot size requirements. There is a nationwide shortage of missing middle housing—duplexes, fourplexes, small apartment buildings—because it's so difficult to find places to build these things.
In this context, try coming up with a novel mix of those attributes people look for in a house or neighborhood. Where can you build a place with that mix? Where can't you?
We also don't unbundle because our development model privileges large increments of capital and very large-scale, master-planned communities. And developers have a template. "Cookie-cutter subdivisions" is a pejorative based in a lot of truth.
Consider the much-derided "snout house," with the garage in front and the "front" door comparatively off to the side and nondescript. Pretty sure no one actually likes these things aesthetically. It's like living in an attractively landscaped loading dock. Developers of this stuff insist, "This is what the market wants," based no doubt on survey research. No, the market would probably love it if you could deliver the convenience of driving (given the presumption of an environment that requires driving) in a different way.
The Village Homes subdivision in Davis, CA does it by having the houses front onto pedestrian paths and cars park in the back at, well, pretty much loading docks. It's a wonderful template I'm shocked hasn't been repeated more often; I love Village Homes and may well write about it another day.
Developers have their template, and it may get tweaked at the margins, but it rarely gets radically rethought. And our zoning codes, our housing finance systems, pretty much everything that could be is weighted against radical experiments, even small ones.
What we need is thousands of small experiments that would give us a truly diverse mix of housing options and neighborhood options. We need the targeted deregulation that would help unleash it. A nation of Strong Towns would be a nation in which our housing choices are far more "unbundled" than they are now. The market could then evolve far more effectively in response to actual consumer preferences, because those actual preferences would be more accurately revealed if we had greater real choice.
All photos by the author unless otherwise indicated.