Parking Requirements are a Mandate for Expensive Housing

February 12, 2021

An abundance of academic literature strongly suggests that parking requirements hurt American households by raising the cost of housing and increasing demand for automobile travel by artificially lowering the cost of car storage. Essentially, by reserving land for parking irrespective of market demand, the marginal propensity to drive increases, but this cost is borne by higher housing prices. A new paper from Gabbe et al (2020) adds to this literature by studying the effects of Seattle’s minimum parking requirements—and what happens when they go away.

Key takeaways:

  1. Parking requirements are the single greatest predictor of parking provision in new developments. Out of 868 new developments studied between 2012-2017, one third provided exactly the number of parking spots required, not one more or less.
  2. 70% of developments with no parking requirements did include some parking. This parking was associated with lower population density, mixed-use development, and higher employment density, but the effect was weaker overall than parking mandates.
  3. In the five years being studied, developers provided 40% fewer parking spaces, saving around $537 million.

If no one requires 18,000 parking spaces, does anyone build them? Gabbe et al (2020) find a resounding “no” in the data. Nevertheless, private market actors will still provide some parking, absent land-use regulations requiring it. The study suggests that parking requirements create a costly glut of parking spaces, misallocating valuable urban land amid a high demand for housing. Earlier research from Gabbe and Pierce (2017) finds each garage parking space can raise average monthly rents by $141. For renters who do not own cars, requiring a parking space with their home effectively forces them to spend that money to consume space they don’t need. But to what extent does removing that requirement reallocate these resources effectively?

The authors set out to answer to basic questions: “First, how do parking requirements affect the
amount of parking provided in new residential developments? Second, how does the response to requirements vary by building and neighborhood characteristics?” A regression analysis of permitting data on 868 new developments in Seattle, spanning from mid-2012 to late 2017, attempts to disentangle the variables. Seattle has been advancing land-use reforms to allow more Transit-Oriented Development in so-called “urban villages” where housing is sorely needed, near frequent transit and employment centers, to reduce commute times and housing costs. In 2012, it eliminated the requirement to provide a minimum of one parking space for every new housing unit built in these urban villages.

For the roughly 60,000 new homes built in the five years since this reform, the authors model the effect of parking requirements by first asking what would have happened otherwise. “ If a completed development included less parking than the pre-reform minimum requirement, our counterfactual is that the developer would have included the pre-reform minimum,” says Gabbe et al. “For example, if a developer (post-reform) included 0.4 parking spaces per unit – but would have been previously subject to a one parking space/ unit minimum – the counterfactual scenario is that the developer would have built one parking space per unit.” The study’s regression models should be able to explain what happened to that hypothetical extra 0.6 that did not materialize.

“Nearly 20% of new residential buildings in our dataset include no parking, and about 88% of buildings contain less than one parking space per unit,” the study finds. “The average building has 0.68 spaces per unit, while the average building in areas with reduced or no parking requirements has 0.57 spaces per unit.” The most a developer built in excess of parking requirements typically did not build more than an additional 0.5 spaces for every unit. In other words, if a given parcel’s zoning required 0.3 spaces per unit, developers almost never built more than 0.8 per unit.

Using the most conservative cost estimates, the researchers project a savings of $537 million in hard construction costs. This savings is thanks to the absence of roughly 18,000 parking spaces that would have otherwise been required. Notably, though, this figure “does not include the opportunity costs associated with building parking in spaces that could be used for housing”—a cost of paramount importance in housing policy.

Because the regression models also find a negative correlation between population density and parking provision, the authors conclude that the tradeoff between space for car storage or human habitation is fairly direct. Although 70% of developments with no parking requirements still provided some parking space, these extra spaces followed eager consumers in some predictable ways. Parking provision was higher in areas with higher land values, “likely to satisfy tenant demand.” In other words, wealthier tenants, including commercial tenants in mixed-use developments or tenants closer to job centers with scarce parking, were willing to pay the extra cost of car storage, so developers provided it.

Eliminating parking requirements can provide a powerful incentive to allocate scarce urban land more efficiently. This study suggests that the effect of this incentive is dramatic and immediate, but it also does not mean cities will abolish all parking overnight.

It’s also important to consider how this reform makes Seattle relatively unique in American urban planning practices. “Lastly,” the authors note, “all but one development had less than two parking spaces per unit, a level of provision that is common (and sometimes required) in other cities.”