Valuing Open Space in a Locational Equilibrium Model of the Twin Cities

Abstract

This research employs recently developed econometric techniques to extend the literature on open space as well as present one of the first micro-level analyses of housing supply. To estimate horizontal sorting models for housing demand and housing supply, a new database consisting of over 460,000 residential transactions spanning 1990 through 2006 as well as detailed land use data identifying nine types of open space is obtained for the seven county Twin Cities metropolitan area in east-central Minnesota. The decision to estimate a horizontal sorting model allows the recovery of preferences for local public goods that vary across observable attributes of both households and builders. Several policy implications arising from the estimation results are that: (i) potential development of open space reduces the amenity value open space provides to households, (ii) different types of open space are associated with very different marginal willingness to pay measures indicating the importance of treating open space as a heterogeneous good, (iii) estimation of heterogeneous preferences shows that both households and builders have heterogeneous preferences and accounting for that heterogeneity provides important information to policymakers, and (iv) policies restricting the amount of developable land can influence the location of the supply of new housing by builders.

The use of a horizontal sorting model to estimate both households' demand for housing and builders' supply of new housing allows analysis of not only marginal willingness to pay measures for local public goods, but also willingness to pay measures associated with non-marginal policy counterfactuals. Integrating both the demand and supply sides of the housing market, I estimate general equilibrium welfare measures where both households and builders adjust their housing decisions in response to policy counterfactuals. To clear the market, I adjust the price of housing in order to remove excess demand and excess supply using an iterative numerical technique. I find that significant differences exist between partial and general equilibrium welfare measures and show that incorporating builder supply responses to policy counterfactuals results in potentially large welfare differences compared with both partial equilibrium measures and general equilibrium measures assuming an exogenous supply of housing.