Hidden Costs and Deadweight Losses: Bundled Parking and Residential Rents in the Metropolitan United States

Abstract

There is a major housing affordability crisis in many American metropolitan areas, particularly for renters. Minimum parking requirements in municipal zoning codes drive up the price of housing, and thus represent an important potential for reform for local policymakers. The relationship between parking and housing prices, however, remains poorly understood. We use national American Housing Survey data and hedonic regression techniques to investigate this relationship. We find that the cost of garage parking to renter households is approximately $1,700 per year, or an additional 17% of a housing unit’s rent. In addition to the magnitude of this transport cost burden being effectively hidden in housing prices, the lack of rental housing without bundled parking imposes a steep cost on carless renters—commonly the lowest income households—who may be paying for parking that they do not need or want. We estimate the direct deadweight loss for carless renters to be $440 million annually. We conclude by suggesting cities reduce or eliminate minimum parking requirements, and allow and encourage landlords to unbundle parking costs from housing costs.

Keywords:

• Affordability
• land use/zoning
• prices
• regulation
• rental housing
• transportation
• urban planning

Acknowledgments

We are grateful to Randy Crane for many interesting discussions about the American Housing Survey and parking policy. We thank Don Shoup, Mike Manville, and attendees of the 2015 Association of Collegiate Schools of Planning conference for their insightful comments on previous versions of this study. We also appreciate the thoughtful feedback of three anonymous reviewers.

Notes

1. Surface parking, on the other hand, is generally provided in areas where the opportunity cost of not building housing is lower.

2. We aggregate county-level data from the 2010 American Community Survey to the 1980 standard metropolitan statistical area definitions used in the 2011 American Housing Survey.

3. Specifically, the survey asks new movers whether they moved to their current residence because of the quality of neighborhood factors including schools, public services, transit, and general appearance. Data on these additional neighborhood variables are available for a limited set of urban rental households (n = 13,677). We include these variables in the same model as above with a more limited sample.

4. All in 2011 dollars. The AHS top-codes rental values at the 99.5th percentile, so we would expect these averages to be marginally higher in the absence of top-coding. However, this does not represent a concern for the integrity of the analysis, as we are most interested in parking’s effect on low- and moderate-cost rental housing, rather than the high end of the rental market.

5. The substantive result of our modeling was unchanged by employing a clustered standard error approach. Although coefficients are reduced and confidence intervals are increased in this specification, the statistical significance and sign of the relationships between individual attributes and rental value are consistent, including the value of garage parking. This gives us further confidence that metropolitan-level effects, although important, are not driving the model results.

6. We lack data on the prevalence of renters subleasing their parking spaces, but even if a small share of renters do rent out their garage parking, the amount of wasted resources is substantial.