A recent paper, Houston, you have a problem: How large cities accommodate more housing, by Anthony W. Orlando and Christian L Redfearn, offers a new reading of real estate data.

Consider the stylized fact that unmet demand is most-inexpensively delivered on low-cost land at the periphery of the commuting shed, known as a “greenfield” site. This type of development uses low-cost, low-density construction methods. However, in productive and desirable urban areas, low-cost land—especially close to jobs and retail—is quickly consumed, pushing single-family home builders farther away from the amenities that make these urban areas attractive. Eventually, this progression reaches a limit in which commuting back to these amenities is too costly. At this point, the greenfield land is effectively “built out,” and developers are forced to look inward to more expensive land closer to the core where spatial amenities are valued by renters and buyers. When this “infill” development becomes a larger share of new housing supply, the marginal cost of supplying a new housing unit will increase, and the elasticity of supply will fall. Thus, even in the absence of different regulatory regimes, an MSA with more population and more density will appear to have a steeper supply curve because large and growing urban markets naturally progress in this direction.

Real estate has a history of being talked about in static numbers. Orlando and Redfearn discover a dynamic in their research. A city grows along the fringe where the developers can build over large parcels of undeveloped land. This is the most consumer-friendly by meeting the desired structure for the lowest cost. But at some point, the authors observe that the commute to a central business district causes infill projects to gain in status. At that point, a city gains new units within the old infrastructure instead of in the greenfield.

Much of what we have learned in the two decades since DiPasquale (1999) first prompted the field to investigate housing supply is aggregate and static in nature. The goal of this empirical work is to document the location of housing stocks within several MSAs over a long time of growth. The results presented in the article are largely descriptive. It is abundantly clear that aggregate analyses miss the compelling dynamics we documented.

Why stop at the trade-off between low cost fringe housing versus commute time? There are many other interesting dynamics to expore.