value may not guarantee the socially efficient level of open space to be self-financed. For
the socially efficient level of open space to be self-financed, the marginal change rate of
post-preservation equilibrium land price with respect to preserved open space must be
less than or equal to-----ɪ----* +--T'-----—, the financially defined marginal
L - a0- a* (L - a0)wτ+ a*δ
cost. Although derived based on spatially homogeneous open space amenity, these
conditions can be extended to spatially distributed open space amenity if people’s bid
price is taken as a spatial average.
Our simulation results for the spatially explicit open space model not only show
the existence of an optimal amount of open space that can be financed by property tax
increment even for a weak preference for open space preservation (with a utility elasticity
of 0.04 with respect to open space), but also illustrate the spatial configuration of open
space does matter in terms of the net value of community developable land and the
capacity of tax increment financing. Generally speaking, an evenly distributed, centrally
located open space can achieve greater net social value and stronger capacity of tax
increment financing than other spatial configurations of open space. That is, a central
location is better than non-central location, several small pieces is better than one large
piece, a ring shape is better than a circle, and a cross shape may or may not be more
efficient than a ring shape. However, a central community park may be politically
desirable by less administration or transaction cost involved in the acquisition of involved
open space land if private ownerships of the land involved are relatively concentrated.
These optimal spatial configurations, we suspect, are very likely to be robust, since they
tend to maximize the coverage of the positive externality of open space. Moreover, the
people’s preference and the description of open space amenity used in our simulation are
36