Public goods and spatial location: stated preference prioritization with spatial interdependencies
Date
2011
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Public goods exist in space and recent research has shown that accounting for the location of their provision can have many implications for policy. Non-market valuation techniques are often utilized to value benefits from the provision of public goods in monetary terms. However, if willingness to pay (WTP) for public goods varies with the spatial location of provision, then valuation studies without spatial specificity may lead to suboptimal provision. WTP for land preservation provides a public-goods context where space matters. This study examines the benefits of preserving farm and forest parcels, which are derived from existing survey data from community-level and state-level choice experiments (CE) conducted in 2005 in Delaware and Connecticut (Johnston and Duke 2007; Duke and Johnston 2010). Data on applicant parcels, both accepted and rejected, to the Delaware Agricultural Lands Preservation Foundation (DALPF) from 1995-2003 (Allen, et al., 2006) are utilized to develop a hedonic model to estimate conservation easement costs across all eligible parcels in Sussex County, Delaware. Using a geographic information system, nonuse benefits and easement costs are predicted for every eligible parcel in the County to cost-effectively prioritize parcel selection. Parcels are selected under a fixed budget with four different prioritization strategies: benefit-targeting (BT), cost-targeting (CT), benefit-cost targeting (BCRT), and an binary linear programming optimization method (OPT). This initial prioritization effort does not account for variability in benefits associated with the spatial location of selected parcels. To investigate potential spatial interdependencies in parcel selection, a distinct 'gravity' is calculated for each parcel i to every other parcel j in a feasible set based on the size of the parcels and the distance amongst them. An exponent (β) on distance captures the friction of these distances (i.e., high β, higher friction). The gravity values are then scaled by α to determine the spatial synergy (SS) benefits for all parcels i and j. Parcels are selected utilizing an algorithm for a quadratic knapsack problem (QKP), which finds a subset of parcels that maximizes net social benefits subject to the constraint of the knapsack—in this case, the budget. Because the true value of α would reflect WTP for spatial proximity and this is unknown, a sensitivity analysis is conducted by varying α to allow for policy makers to evaluate the level at which spatial preference might affect optimal choice. Results from the County-level model demonstrate that optimal (OPT) and near-optimal (BCRT) parcel prioritizations both generate $435 million more net benefits to society compared to BT and nearly $236 million more net benefits than CT. Furthermore, accounting for spatial interdependence dramatically alters the optimal set of parcels selected and thus the preserved landscape. Any study ignoring space, when public preferences vary systematically with space, will thus produce a suboptimal provision of preservation. The sensitivity analysis shows the degree to which this spatial preference varies with changes in the hypothesized values of α and β. This outcome indicates that welfare analysis based on a standard valuation study might, in fact, provide information that leads to misguided policy. Since the primary land preservation decision is whether to preserve a specific parcel, consideration of systematic elements, such as spatial interdependence, is needed to avoid potentially arriving at suboptimal policy guidance.