Landscape Preferences and Patterns of Residential Development

I analyze a model of residential location choice in which people derive utility from their proximity to open space. When people have such landscape preferences a new residential development contains more people, more tightly packed than is optimal. More surprising, in a model where new residents arrive simultaneously, I find that land price gradients are highly non-monotonic and do correctly reflect the value of open space. On the other hand, when new residents arrive sequentially, land price gradients are nearly monotonic but do not correctly reflect the value of open space. Finally, dynamic equilibria generally have the property that more remote areas are developed before more central areas. These results have a number of interesting implications for policy. In particular; (1) the creation of central city parks is welfare improving, (2) infill development of central city open space is not welfare improving, (3) the ability of regulation to restrict development at the city s limit, greenbelts , to improve welfare does not derive from a taste for nearby open space, and (4), creating small parks in undeveloped areas before they are subject to development pressure may deter leapfrogging development. Finally, the fact that land prices need not reflect the value of open space suggests that hedonic estimates may understate the value of such open space.