Sprawl and the Cost of Living

Cross-posted from City Observatory

Over the past three weeks, we’ve introduced the “sprawl tax”—showing how much more Americans pay in time and money because of sprawling urban development patterns. We’ve also shown how much higher the sprawl tax is in the US than in other economically prosperous countries, and how sprawl and long commutes impose a psychological, as well as an economic burden. Today, we’ll take a close look at how ignoring the sprawl tax distorts our view of the cost of living in different regions and neighborhoods.

As one old saying goes, an economist is someone who knows the price of everything and the value of nothing. It’s often claimed that some places, often sprawling Sunbelt cities, have a lower cost of living, based usually on observations about lower housing prices. And judged solely from the sticker price of new homes, the argument has some merit.

Phoenix. Credit: Al_HikesAZ, Flickr
Phoenix. Credit: Al_HikesAZ, Flickr

But as our aphorism about economists implies, there is a lot more to this question than just one set of prices. If you’ve followed our series on the sprawl tax, you know that living in some cities—those with cheap average housing costs, like Houston or Dallas or Birmingham—also carries with it a heavy, and largely ignored cost in the form of the “sprawl tax”: much higher transportation costs. In short, we tend to fixate on the price of something we can easily measure (housing) and simply leave out the value of something that is much less obvious (sprawl and longer commutes).

How big is the sprawl tax, relative to the supposed cost of living differences among metropolitan areas? Quite large as it turns out: enough to erase much of the supposed cost advantages that low density settlement is supposed to offer.

We know that within metropolitan areas, there’s a strong tradeoff between rents and home prices and typical commute distances. Low density housing, at a long remove from the city center and most metro area jobs, commands lower prices, reflecting in large part the added transportation costs implied by more far-flung and less accessible locations. Others, notably the Center for Neighborhood Technology through its H+T calculator, have addressed the tradeoff between housing costs and transportation costs within metropolitan areas. Our analysis here uses the sprawl tax in concert with Bureau of Economic Analysis estimates of housing costs differences at the metropolitan level to examine this tradeoff.

The Sprawl Tax and Cost of Living Differences

In the following chart, the vertical axis shows the sprawl tax: higher values mean that a region’s residents pay more for transportation costs and travel time as a result of sprawl. The horizontal axis shows the rent differential: how much more (or less) the typical resident pays in annual rent/housing costs compared to the typical large metropolitan area. Areas to the right have higher rents; areas to the left have lower rents. And, as discussed in the post-script, at least some portion of the difference in rents reflects real quality of life differences between cities; but for now we use this as an index for housing cost comparisons.

The chart illustrates a range of different combinations. Four metro areas with the highest sprawl taxes (Atlanta, Nashville, Dallas, Houston) all have lower than average housing costs. For example, according to the BEA estimates, the annual cost of housing in Houston is $850 per person less than the national average. But the typical Houston household would face a sprawl tax based on longer commute distances of about $2,900 per worker which would essentially wipe out the superficial price advantage from housing.

The nation’s most expensive housing markets—including San Francisco, San Jose, New York, and Washington—have housing costs that are considerably higher than the national average, with per person annual rental costs exceeding the national average by $5,000, up to almost $10,000 in San Jose and San Francisco. But these cities have much lower than average sprawl taxes. Workers in San Francisco, San Jose and New York pay less than $500 per year in sprawl taxes. This amount doesn’t fully offset the added rental costs, but combined with the quality of life differences between high prices and low prices cities, makes the differences much smaller.

The Bronx, NYC. Credit: Dave Johnson, Flickr
The Bronx, NYC. Credit: Dave Johnson, Flickr

And for many cities, adding the sprawl tax essentially erases the supposed cost of living advantage from cheaper housing. For example, Houston’s housing cost is about $1,100 per person less than Portland’s ($845 below the large metro average compared to $318 above). But Portland’s sprawl tax ($871) is $2,000 less, per worker, than Houston’s ($2,877), which for many households will more than eliminate the housing price advantage.

A sensible discussion of the real differences in the cost of living between places has to look just past the prices of a few commodities and big ticket items, to consider the intrinsic value consumers attach to great urban environments, the variety and convenience of consumption opportunities in compact urban centers and the sprawl tax the consumers must pay.

Methodology: How we calculated living cost differentials

The best measure we have of inter-metropolitan differences in consumer prices is the Bureau of Economic Analysis Regional Price Parities (RPP) estimates. These estimates, prepared from the data used to construct the nationwide consumer price index, confirm some of our fundamental intuitions about differences in costs of living between communities. The cost of goods varies little among places. The difference between relatively expensive and relatively inexpensive cities (that is, the 75th percentile and 25th percentile) in the cost of goods is just 2.5 percent. For services, prices vary somewhat more, about 8 percent. So what’s really making up the difference in cost of living? No surprise: housing, which varies by more than 30 percent.

Regional Price Parities for Metropolitan Areas of Over 1 Million, 2013

Component Weight Mean Interquartile Range (25%-75%)
Goods 41.5% 99.2 97.3 – 99.7
Services (excluding rent) 37.9% 99.3 94.3 – 102.1
Rent 20.7% 107.2 85.5 – 116.0
Source: Bureau of Economic Analysis, City Observatory calculations. Weight is the share of these items in the RPP calculation. Index values are for the entire nation, i.e. U.S.=100. Mean and interquartile range are weighted by metro area.

Here are BEA’s estimates of the rent price index for the 51 largest U.S. metropolitan areas (those with a population of one million or more). The data are indexed to the national average rent (U.S. = 100). On average, rents are about seven percent higher in large metropolitan areas (i.e. the index for the median metropolitan areas of the 51 largest is 107). Rental prices in San Jose and San Francisco are 80 to 90 percent higher than nationally; rents in Louisville and Birmingham are 25 to 30 percent lower than the national average.

Screen Shot 2016-07-06 at 12.46.39 PM

Because housing costs are the biggest source of variation in the measured cost of living among large metropolitan areas, we use the BEA data to estimate how much more, or less, a typical household pays for housing, based on the difference in rents among metropolitan areas. BEA estimates the rents as a combination of the actual rent paid by renters, and the “imputed rent” which is the value of housing services received by households that own their own homes). We compute the difference in the income paid for housing among metropolitan areas by observing the difference between the average rental price parity for the 51 largest metropolitan areas (107) and the actual value for each metropolitan area. We multiply that difference by the per capita personal income of the area and the share of per capita income devoted to rent (estimated by BEA at 20.7%).

A Postscript: Two Big Challenges with Comparing Living Costs

We present the estimates of housing cost differentials here, and compare them to the magnitude of the sprawl tax to illustrate how important urban form is to our economic and personal well-being. But its also worth noting that conventional cost comparison measures understate two important economic advantages of cities.

First, as with other commodities, differences in prices often signal the value that consumers attach to different objects. Housing may be more expensive in San Francisco or Hawaii than in Omaha or Idaho, but much of this difference reflects the value that we attach to being in a vibrant city or a sunny, tropical climate. A two bedroom apartment near the beach in Maui or on Nob Hill in San Francisco represents an entirely different set of amenities than a similarly sized apartment in Fresno or Fargo. Indeed, a whole class of economic analysis—hedonic regression—uses these price differences to estimate the value of natural and manmade amenities.

Second, there’s an inherent limitation in trying to compare different places that have widely different sets of attributes and consumption opportunities. Typical estimates of cost-of-living differences will look only at a single commodity (like housing) or a very limited market basket of simple goods and services, and use these to compute differences in living costs between places. The assumption is that consumers or households buy the exact same mix and quantity of goods and services wherever they live. Economists have cast serious doubt on the validity of these simple-minded price comparisons. It turns out, consumers attach value to having convening access to a big range of goods and services, something you find mostly in cities. Columbia’s Jessie Handbury has shown that the greater mix, variety and convenience of shopping opportunities in larger cities means that consumers actually enjoy lower prices for the particular market basket of goods they prefer in larger cities than in smaller ones, contrary to the notion that small town prices are lower. BEA’s Regional Price Parities which imply that New York’s prices for goods are 8.8 percent higher than the national average aren’t adjusted for quality and variety. According to Handbury’s estimates, when you make that adjustment, prices in larger cities are actually lower than in smaller ones.

67 thoughts on Sprawl and the Cost of Living

  1. Other reasons trains are useful:

    1) Less impact on the natural environment in terms of land area used for transportation.

    2) Much higher potential speeds up to 3 times as fast as driving.

    3) They enable those have can’t afford a car or can’t drive, or just don’t want the hassle owning/maintaining/driving your own vehicle, to travel.

    As for the vastly expanded transportation demand automobiles supposed enabled, that was precisely because we built lots of sprawl which was inaccessible otherwise. 60 years ago there was nothing worth going to in these areas. Now people travel to them to see friends or family who live there, or commute to work, by car. Had people remained in cities, the primary reason for auto travel to be commonplace largely wouldn’t have existed. Without that, you wouldn’t have had economies of scale for roads or cars themselves.

    But even in Europe, with its high-density cities and high-quality well-funded transit systems, far more passenger-km happen by automobile than transit.

    Passenger-km skews things in favor of automobiles. We should look at the number of trips instead. Also, regardless of which metric you use, the percentages who use automobiles regularly in Europe are far less thanks to their public transit systems. If people have cars at all, they’re often only used for weekend trips outside the city, not daily commuting to work. It’s actually possible to live in lots of places there other than just large cities without owning a car. That’s how it should be in the US. Car ownership should be an option for those who want it, not a necessity just to get to work.

    You aren’t seriously suggesting we build a transportation system based on f**in’ crazy bikers drafting buses, are you?

    No, but velomobiles give you most of the advantages of drafting. A strong rider in one can maintain 35 to 45 mph. A transportation system partly based on velomobiles wouldn’t be a horrible idea. They have most of the speed/convenience advantages of cars without the other hassles. No licensing, no insurance, no fuel, etc.

  2. It’s a chicken or egg thing. If more people asked for bike accommodations on public transit they would exist. Also, it isn’t always about getting to your destination in the least possible time. If going part way by bike adds a few hours to what is already a 24+ hour train trip it’s not a huge deal if it means you’ve saved the cost of owning a car. That can pay for lots of trips you otherwise may never have made.

    Anyway, why the glory in the train? They are far less environmentally friendly (compared to a Prius) than we’ve been led to believe.

    You need to look at all the energy costs, which include the embedded costs of the vehicles themselves, the cost of the guideway, the costs of securing energy supplies, policing, courts, etc.

    Even if trains come out even with econoboxes carrying a full passenger load, they still come out ahead in terms of space, passenger comfort, and speed. You’re not going 200 mph in a Prius but you can travel at those speeds on a train. A cramped Prius is a really awful way to travel for hours. Really, any car is. The largest car is still far less comfortable than a train. Train is just a higher quality option all around.

    I looked at things logically once and concluded Americans have been sold a bill of goods. There are few transportation niches where cars are the best answer. In cities they totally suck. They’re slow, space intensive, and make life miserable for everyone outside of them. In the suburbs rail works better for work commutes to the cities, and bike can work fine for local trips which are often just a few miles. For long distance trips plane or train is always better. From where I stand the only place automobiles have a niche is in rural areas where other forms of transport just aren’t viable. It’s amazing they caught on to the extent they did considering their cost and extensive licensing/insurance requirements.

  3. If you have trouble falling asleep within 50 miles of NYC, and think the sky is dark there, I’d recommend you expand your horizons.

    Where you were is known as suburbs (pretty dense ones too, by US standards), and you’re right, it is pretty nice there (maybe you can see why some would like to live there), but that’s nowhere near nature, unless your only baseline is central park.

    As for the sky; heres the dark sky map:

    As you can see, you have to go pretty far (Approx 130 miles out from midtown) to see a dark sky (light green or better).

    I’ll argue that you just don’t know what you’re missing.

  4. That’s exactly my point; cars can give you a different perspective, if you have no issue with a 4 hour bike ride, a 3 hour drive (~150 miles) would be a piece of cake.

    I’m glad you mentioned stargazing, since it’s something I hold quite dear. Back before I drove, seeing the milky way was a once a year thing, if that. Now I can go see it whenever I want without too much hassle as long as the weather cooperates.

    And believe me, seeing the milky way never gets old 🙂

  5. The US national railway passenger system (such as it is) is Amtrak–lose the ‘c’.

  6. I don’t think there are any boats that go out that far (~40 miles to get to a green zone) on a regular schedule. Chartering a boat to go that far is rather expensive, and you have to deal with often harsh wind and waves. Not to mention somebody else’s schedule.

    Have you done it? it’s not exactly a piece of cake. Not to mention that telescopes are hard to use when what you’re standing on is moving.

    As for sails… uh, have you sailed before? A twenty-something-foot sailboat has a hull speed of maybe 7 knots… Assuming you don’t need to tack at all… you’re going to spend 5 hours just going out far enough. Hope you brought snacks!

  7. Once the LED street lights are in place everywhere here in the northeast, there’s a middling chance you might be able to see the Milky Way at least from parts of the outer boroughs. Certainly you’ll be able to see a lot more than you already do. I’m already starting to notice darker skies even though the LED streetlight installation is far from complete.

  8. Simply put, no. I’m out on LI pretty often, where most places have switched over to LEDs already, and although it’s definitely darker, it’s not even in the same ballpark as what you can see when you’re far away. It takes a pretty darn dark sky to see the milky way.

  9. For what it’s worth, LI still has a lot of light pollution from NYC. It’s amazing how far it carries. A good start besides the LED street lights would be to get businesses to turn off the lights in office buildings at night. It’s a pointless waste of energy also keeping lights on when the buildings are empty. We could also start mandating dark sky friendly lighting for parking lots and other commercial lighting.

  10. It works both ways. “imply that New York’s prices for goods are 8.8 percent higher than the national average aren’t adjusted for quality and variety. “. The notion of the authors “Sprawl Tax” also, does not factor in the adjustment for quality & variety, in than many people prefer to live in the suburbs. Our existence isn’t defined by transportation elements. The myth that high density living in urban areas is somehow a higher quality lifestyle is very much in dispute.

  11. Yeah so, looking into it, I realized a couple of things. People probably generally don’t move based on taxes, with the exception of retirement, but poor transient people may benefit from living in a state with no income tax rather than an EITC, even a generous one as New York City’s is, which they’re not going to know how to file for, or the added step of filing income tax and getting a regular refund, etc.

    Leaving that aside, let’s assume a person or family not on some kind of direct tax rebate caused by low income is in New York City. Let’s keep it simple and say a single renter making $30,000, and I’m going to leave out health insurance deductions etc. They’re paying an 8.875% sales tax, one of the highest in the country, on most items, and they’re paying the highest income tax of anyone at their income level in the country. According to https://smartasset.com/taxes/income-taxes , that person would pay about $1800, 6% of her income, in income tax in NYC (this probably does not factor in the $75 school tax credit so let’s say $1725). OK, this same person in California would pay $588, a little less than 2%. In Texas this person would pay zero. None of these people likely has enough deductions to itemize, but the difference between low tax-obsessed Texas and leeeberal California is actually minimal (NY has a very small SDI which may be imposed and can be ignored here, California’s is 1% I think. So let’s say actually $880 in CA tax/SDI).

    So what? Well, if we doubled her income to $60,000, in California she’d pay $2680, about 4.5% (5.5% w/SDI). In NYC, $4,800, or 8% (minus school tax credit $75 less). This is an effective tax rate, not marginal. In Texas, again zero. And in major Texas cities, an 8.25% sales tax rate; slightly lower than NYC. Again, this person does not have enough to itemize. The feds are not helping.

    But let’s triple it, to $90,000. Now New York City starts to not look as bad; the effective tax rate only goes up to 8.7%, and is $7,768 after the school tax credit. This person can itemize and, as she’s in the federal 25% bracket, the feds are paying 1/4 of every dollar in state income tax paid after the first $6300. By my quick calculation here she’s paying an effective 8.2% after the deduction. She’s still paying high taxes, but the tax isn’t actually that progressive at this point.

    In California, 6.08%, $5,470. Not enough by itself to itemizable, but she’s in the ballpark of deductibility with the 1% SDI. It’s still a full 1/4 lower than NYC, and sales tax is a little lower also. In Texas, still zero.

    My point is, if you’re doing the median $30K in America, New York City is tough. But, you have the fixed, low cost of transit and benefits of relative walk/bike-ability to help you out. If you can find an affordable housing situation, you can be sitting relatively pretty.
    If you’re doing well, at 60K, you’re absolutely getting hosed on taxes, and transit is not offsetting that anymore. This is where the sprawl tax case falls way short.

    But if you almost make it to six figures, you’re paying a lot of taxes- but a barely higher share than before. The marginal value of a dollar to you has decreased, and it’s much more of a personal value call, because the tax system, to some degree, is actually working for you here.

    (what happens if you live in New York City and work in Yonkers? Do you pay both local income taxes on top of state income tax? Such people would be the highest-taxed at that income in the US)

  12. why use only one transportation form? why not mix & match and share & combine different vehicles such as bus, train, private car (especially with bike carrier) , uber even trailer.? lets say i have a car and a bike. i could share said car with one or two others who having a bike and/or motor scooter themselves would not need the car all the time and i wouldn’t either if i have a bike.//// say two of us work at places a mile apart. i can drive my car to a friends house or he can ride his bike to mine and put his bike on my carrier. i can them drive to work and drop my friend off at an opportune place and he can bike to his work place from there. if said friend leaves work later or earlier than me he can bike home directly or to a bus stop where he can put his bike on the bus to do so. in a similar way i can share my friends scooter if we work at the same place or nearby.///bike trailers or motorcycle trailers can be to put in this mix. their hitches can be designed to make them attachable to cars or bike trailers attachable to motorcycles or motorcycle trailers attachable to bikes. why can”t i pull an adult in a bike trailer? roller blades can provide access to mass transit. and etc.


Leave a Reply

Your email address will not be published. Required fields are marked *


Study: Annual Cost of Sprawl in America Adds Up to $4,500 Per Person

A new study confirms what we already know too well: Sprawl is expensive. The New Climate Economy’s latest report [PDF] attempts to put a figure on it and it’s pretty staggering: more than $1 trillion a year nationwide. That figure accounts for higher costs to individuals and communities associated with sprawling development patterns, including transportation infrastructure, less […]

Introducing the Sprawl Tax

Crossposted from City Observatory.  If you read the news, you’ve probably seen reports about “congestion costs”: how much American commuters pay, in money and time, when they’re stuck in traffic. It’s fair to say that we’ve got some issues with many of these reports—but they’re popular nonetheless, perhaps because they help quantify a frustration that […]

Lousy Neighborhoods, Not Lax Zoning, Make Sunbelt Houses Cheaper

The middle class is getting priced out of liberal cities, while red-state urban areas remain affordable. Does that mean our cities should be less like tightly regulated San Francisco and more like permissive Houston? It’s a common argument — but it doesn’t fit the facts. To start with, Houston is hardly a paradise of deregulation. In practice, local experts explain, […]

How Sprawl Got Detroit Into This Mess

It wasn’t de-industrialization that bankrupted Detroit, wrote Paul Krugman in a New York Times column yesterday. If that was all there is to it, then how do you explain the fact that Pittsburgh, once so dependent on the steel industry, is now recovering? No, what brought Detroit to this low point, more than the loss […]