The urban economics of sprawl

Should YIMBYs support or oppose greenfield growth? Two basic values animate most YIMBYs: housing affordability and urbanism. Sprawl puts those values into tension.

Let’s take as a given that sprawl is “bad” urbanism, mediocre at best. Realistically, it’s rarely going to be transit-oriented, highly walkable, or architecturally profound.

So the question is whether outward, greenfield growth is necessary to achieve affordability. And the answer from urban economics is yes. You can’t get far in making a city affordable without letting it grow outward.

Model 1: All hands on deck

Let’s start with a nonspatial model where people demand housing space and it’s provided by both existing and new housing. Existing housing doesn’t easily disappear, so the supply curve is kinked.

A citywide supply curve is the sum of a million little property-level supply curves. We can split it into two groups: infill and greenfield, which we add horizontally.

If demand rises to the new purple line, you can see that the equilibrium point where both infill & greenfield are active is at a lower price & higher quantity than the infill-only line. The only way to get some infill growth to replace some greenfield growth, in this model, is to raise the overall price level. And even then, the replacement is less than 1-for-1.

Of course, this is just a core YIMBY idea reversed! In most U.S. cities, greenfield growth has been allowed and infill growth sharply constrained, so that prices are higher, total growth is lower, and greenfield growth is higher than if infill were also allowed.

At the most basic level, greenfield growth is simply one of the ways to meet demand. With fewer pumps working, you’ll drain less of the flood.

Model 2: Paying for what you demolish

Now let’s look at a spatial model where people will pay more per square foot when they are closer to downtown. (If the jobs are evenly dispersed everywhere, the place with the best job access is…the center. So the math is the same for a job-seeker.)

Of course, builders won’t supply new housing unless the price comes in above their construction cost. In general, more intensive types of building cost more per square foot. And building on large greenfield sites costs less (both via scale efficiencies and ease of access/staging) than small-site, infill growth. Finally, infill growth usually has to replace an existing land use.

Taking a specific transect across the city, let’s suppose that the existing value of parcels* as built is shown by the brown line.

[*existing value here has to be expressed in dollars per hypothetical square foot of the parcel as redeveloped. This is weird, but not hard to calculate in a model.]

To generate growth, the cost of development has to fit between the black willingness-to-pay line and the brown existing-value line. And infill typologies are often more expensive, per square foot, than greenfield ones.

What will infill housing cost? First, it will cost more than housing at the fringe, but that’s OK – the people at the core are benefiting in terms of commute time and job access. Paying more to be close to downtown is like paying more for a bigger house – you can’t complain.

Second, infill housing won’t be able to push the price per square foot of new construction below the cost of construction plus the value of existing development. As infill development runs out of abandoned warehouses and parking lots, that wedge will rise.

A successful YIMBY strategy via infill would run into a shrinking gap between falling prices and a rising existing-value wedge. If it got that far, I’d actually be surprised and impressed – but the point of this second model is that the price floor is higher.

Model 3: The tyranny of circles

What neither of the first two models could really tell us is how much greenfield matters as a quantitative matter. The first model shows how all supply channels are additive, the second shows that infill has a higher price floor. The final boss is the circle.

Back in model 2, it looked like there was about as much room for new development at that sharp dip just to the left of downtown (an old warehouse district, say) as in the greenfield areas at the edges. But cities don’t exist on a line – they usually approximate a circle.

In this map of DC, there’s a circle about 2 miles from the center and another about 20 miles from the center. How much more land does the outer one contain?

You don’t need pi for the answer: the outer circle is 100 times bigger in area. And it’s not even at the fringe in most places.

In most debates about the merits of the infill and greenfield, there’s an implicit “acres to acres” comparison – a parcel here versus a parcel there. But that misses the forest: the power of greenfield development is the incredible amount of land that is available for relatively low-cost conversion.

Counterarguments

Don’t greenfield infrastructure costs make it more expensive? If we didn’t subsidize sprawl, wouldn’t it mostly stop?

There are certainly gains from re-using existing urban infrastructure. But the dominance of horizontal growth across civilizations using a huge range of transportation technologies, taxation schemes, regulatory approaches, and infrastructure norms belies the idea that horizontal growth is mostly a subsidy phenomenon.

Cities also come with significant infrastructure costs that intensify with built and human density. I’m skeptical that a full, accurate accounting of costs can be done at the micro level. And the macro-level indicators, like overall tax rates, certainly don’t suggest large savings from density.

Can’t we build greenfield development that’s fully urban and transit-oriented?

There’s a huge variation in the quality of greenfield growth, and we can all applaud developers who are doing interesting things and building suburbs that have the bones for continued growth. But in American cities, transit is almost utterly useless to people living at the edge. It’s the tyranny of circles again: How many transit lines would need to reach the 20-mile radius perimeter to put everyone within a mile of transit? And how long will those commutes take?

Can we allow housing at the urban fringe but ban new highways?

This is just a recipe for stroads. Places like Northern Virginia that failed to build a network of limited-access highways instead have 6-lane arterials with traffic lights that mean they always run at half capacity. For a major city with serious demand pressure, building new highways or parkways is a good and necessary part of greenfield growth.

This is just part of a bigger principle for greenfield growth: Your suburb is never the last one. Much U.S. sprawl is built badly in the specific way that every town, every subdivision pretends that it will always be the last thing before farms and forests.

Yes In My Greenfield

The YIMBY movement is and should remain focused on re-legalizing infill growth. Strong Towns offers good principles for fiscal responsibility in moderate-demand places. There’s room for another policy movement focused on greenfield growth – getting the financing and governance right, planning for future growth, learning from successes and failures, promoting quality design that’s cheap enough for low-end development, and so on. If and when that movement arises, YIMBYs should greet it as a friend and ally: both are important to a future where housing is affordable.