摘要：This dissertation examines the impacts of distance-dependent spatial externalities on patterns of economic activity in a free-market setting. This class of externalities, which include such examples as smog dispersal, pesticide drift, and habitat degradation from roads, are referred to as ``edge-effect externalities''. Under edge-effect externalities, economic optimality will require not only the correct allocation of land to different uses, but also the correct arrangement of land uses. However, an unregulated free market will potentially fail to achieve an efficient arrangement of land uses. Chapter 2 develops a spatially continuous one-dimensional model of edge-effect externalities. The model demonstrates that, while the externality creates an incentive for a recipient to distance himself from the generator, this distance is too small from a social standpoint. The model also demonstrates the potential for positive externalities between those impacted by the edge-effect externality. Chapter 3 formally demonstrates the potential for edge-effect externalities to create non-convexities in the production possibilities frontier. Further, it demonstrates that conflicting border per unit area is a summary measure of landscape efficiency under edge-effect externalities, but this ratio will vary with the number, shape, and geographic concentration of parcels in the externality-receiving use. Chapter 4 develops a two dimensional agent-based cellular automaton model of free-market land use in an economy impacted by edge-effect externalities. It demonstrates that in an unregulated free-market without bargaining, both Pareto-efficient and inefficient equilibrium landscape patterns are possible. Initial configurations of firms, permanent geographic features, and transportation costs will impact final outcomes. Chapter 5 tests the hypothesis that production patterns for California Certified Organic Farms reflect possible avoidance of negative spatial spillovers from surrounding conventional farms. Differences in parcel size, shape, and surroundings between C.C.O.F. and non-C.C.O.F. parcels are demonstrated. While inherently more vulnerable to losses from mandatory buffer zones, C.C.O.F. parcels are shown to potentially lose a much lower proportion of their land to buffers than non-C.C.O.F. parcels. However, very few C.C.O.F. farms border C.C.O.F. farms under separate management, indicating that growers have not managed to coordinate to capture potential positive externalities.
